<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Eniola Ajala on Medium]]></title>
        <description><![CDATA[Stories by Eniola Ajala on Medium]]></description>
        <link>https://medium.com/@eniola-ajala?source=rss-e96e4d3222c2------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*zQqrBK29BzdelC2eR2Fktg.jpeg</url>
            <title>Stories by Eniola Ajala on Medium</title>
            <link>https://medium.com/@eniola-ajala?source=rss-e96e4d3222c2------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 23:48:15 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@eniola-ajala/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[How To Read Research Papers Faster (Without Feeling Overwhelmed)]]></title>
            <link>https://medium.com/@eniola-ajala/how-to-read-research-papers-faster-without-feeling-overwhelmed-ad864295b31f?source=rss-e96e4d3222c2------2</link>
            <guid isPermaLink="false">https://medium.com/p/ad864295b31f</guid>
            <category><![CDATA[reading-research-paper]]></category>
            <category><![CDATA[scientific-research]]></category>
            <category><![CDATA[dissertation-writing]]></category>
            <category><![CDATA[literature-review]]></category>
            <category><![CDATA[reading]]></category>
            <dc:creator><![CDATA[Eniola Ajala]]></dc:creator>
            <pubDate>Mon, 08 Jun 2026 10:56:30 GMT</pubDate>
            <atom:updated>2026-06-08T10:56:30.641Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/694/1*dAQU6SOvgH4z5LSp_BvYCA.jpeg" /></figure><p>If you are a student, researcher, or currently working on a paper or dissertation, chances are your reading list is growing faster than you can keep up with it.</p><p>Every day there is more information to consume, more papers to save, more tabs to open, and more literature to review. You save papers for later. You open dozens of tabs. You promise yourself you’ll read everything over the weekend. Then the weekend arrives, and those papers remain untouched while the feeling of overwhelm continues to grow.</p><p>I will be writing about reading papers fast because if you’re anything like me, you probably keep adding papers left and right to your reading list. Then the weekend comes, and you tell yourself you’re finally going to get through all of them. But somehow, the papers stay unread and the feeling of overwhelm just keeps growing.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/480/1*dlRnHrYLLPQBhAVp1FMwkg.gif" /></figure><p>For a long time, I thought the solution was for me to read more, but I eventually realised that the problem wasn’t the number of papers I had to read but the way I was reading them.</p><p>Like many students, I assumed every paper had to be read from the introduction all the way to the conclusion, line by line. It was slow, exhausting, and often left me feeling like I had spent hours reading without actually retaining much information.</p><p>Over time, I found a few methods and habits that helped me get through papers much faster, remember them better, and actually enjoy the process. So today I want to take you through exactly what helped me.</p><h3>What Are We Actually Trying to Achieve When Reading Papers?</h3><p>I think when we’re reading research papers there are usually three things that we want.</p><p>First, we want to read a lot of papers so that we can keep up with the literature in our field. Second, we want to remember what we’ve read because otherwise all that reading doesn’t really have much use. And third, we want to put those papers into the bigger context of the research literature and understand how they connect to one another.</p><p>Those goals changed how I think about reading research papers. I realised something that changed the way I approached academic literature;</p><blockquote><strong>Not every paper deserves the same amount of attention.</strong></blockquote><p>Once I understood that, reading became faster, more effective, and far less stressful. Here’s the system that helped me read more papers, remember them better, and make sense of the literature without feeling overwhelmed.</p><h3>Start With Your WHY</h3><p>Before opening any paper, ask yourself one question; <em>Why am I reading this?</em><strong> </strong>This might sound obvious, but it’s one of the most important steps. Whenever you pick up a research paper you need to know why. Define your why. Why are you reading this paper? What do you want to get out of this paper? Why is it on your desk or computer? Are you trying to understand the current state of knowledge? Are you trying to get information to support a point you’re making in your writing? Are you conducting a literature review? Are you trying to critique a paper? Are you looking for references to other papers?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/480/1*YcpvzSkoTA4rnpzGTb2Pcg.gif" /></figure><p>You might be reading a paper to understand the current state of research in a particular field, support an argument in your dissertation, conduct a literature review, learn a specific methodology, critique a study, or identify references for further reading. Understanding your purpose from the outset helps determine how deeply you need to engage with the paper and what information you should focus on extracting.</p><p>Many different reasons exist for reading research papers, and you need to know your reason because that determines how you should read it. And also, there isn’t just one way of reading a research paper. There are multiple ways.</p><p>Each of these goals requires a different level of engagement. If you’re only looking for a supporting citation, you probably don’t need to spend an hour analysing every intricate detail. On the other hand, if you’re planning to replicate a study or critically evaluate it, you’ll need a much deeper understanding. Knowing your purpose helps you decide how much time the paper deserves.</p><h3>Quickly Decide Whether the Paper Is Worth Reading</h3><p>One of the biggest mistakes students make is assuming every paper deserves a full read. The first thing I do when I open a new paper is look at the title and abstract. I want to decide very quickly whether it’s actually worth my time.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/480/1*wQ-zA5afr3jCc4wy4H-NNQ.gif" /></figure><p>A lot of papers seem relevant at first glance but aren’t actually that relevant to the question you’re trying to answer. It’s your job as a researcher to decide quite quickly whether a paper deserves your time and effort. I skim the title and abstract, look at the type of results being presented, and then decide whether I want to continue reading.</p><p>Many papers can be ruled out within a few minutes. Once I’ve decided it’s worth my time, then I move to the next stage.</p><h3>Why I Look at the Figures Before Anything Else</h3><p>One habit that significantly improved my reading speed was learning to pay close attention to figures and tables.</p><p>After reading the abstract, I often jump straight to the figures. In fact, one thing that surprises people is that I usually don’t read anything else first. I go straight to the figures. Because figures tell the story of the paper. The figures already tell you a huge amount about the paper.</p><p>By examining the figures, you can often see the kind of results the researchers obtained, how those results are presented, the conclusions that may be drawn from the data, as well as any trends, patterns, comparisons between groups, and the key findings of the study. In many cases, the figures alone provide a clear overview of the paper’s main message before you even begin reading the detailed results section.</p><p>Of course, this depends on how much knowledge you already have in the field. If you’re new to a topic, you’ll probably need to read more of the introduction and results sections. As your expertise grows, you’ll be amazed by how much information you can extract from figures before reading any of the accompanying text.</p><p>After going through the figures, I may go back and look at the methods or results sections if I need additional detail. Sometimes I can understand most of a paper’s findings before reading the results section simply by studying the figures and captions.</p><h3>Never Skip the Discussion Section</h3><p>After examining the figures, I usually jump straight to the discussion section. Personally, I think this is often the most valuable part of the paper.</p><p>The discussion section explains why the findings are important, how the results compare with previous studies, what limitations may have affected the research, and which areas future studies should explore. It helps place the findings within the broader body of knowledge and often highlights opportunities for further investigation.</p><p>Researchers often point out critical limitations, unanswered questions, and areas that need further investigation. Reading these sections can help you think of research questions yourself.</p><p>In some cases, the future work section can inspire the next paper you write or even your dissertation topic. For students and early-career researchers, this section is particularly valuable because researchers frequently point out gaps in knowledge. Those gaps can become future dissertation topics, research projects, or even entire research careers.</p><h3>Create a System for Remembering What You Read</h3><p>Reading papers is only useful if you can find and remember them later.</p><p>To keep track of the papers you read, it can be helpful to use a tool such as Notion. Creating a dedicated workspace for academic papers allows you to organise and retrieve information more easily. Whenever you come across a useful paper, you can use the Notion web clipper to save it directly to your workspace. For each paper, consider recording important details such as the paper link, key findings, significant figures, personal observations, and ideas for future use.</p><p>As you read, you may also find it useful to take screenshots of important figures and add them to your notes for quick reference later. Another effective strategy is to record voice notes after reading a paper.</p><p>Voice notes can encourage more critical thinking because they require you to interpret and explain the information rather than simply copy it. While typed notes sometimes become a repetition of the paper’s content, speaking about the paper forces you to process the ideas, form your own opinions, and identify what you found particularly useful or unconvincing. This approach can significantly improve both comprehension and long-term retention of the material.</p><h3>Put Papers Into a Bigger Context</h3><p>One challenge many new researchers face is seeing papers as isolated pieces of information. Every paper exists within a much larger scientific conversation.</p><p>Understanding those connections is often just as important as understanding the individual study itself. When you’re starting out, one challenge is seeing how papers connect to one another.</p><p>Another useful tool for organising and understanding academic literature is Research Rabbit. Research Rabbit helps visualise how papers, authors, and ideas are connected within a field. It allows you to explore relationships between studies based on publication year, citation networks, research themes, and author collaborations. By presenting these connections in an interactive, tree-like format, the platform makes it easier to see how knowledge has evolved over time and how individual studies fit into the broader research landscape. This can be particularly valuable when conducting a literature review or identifying influential papers and emerging research trends.</p><h3>Use a Reference Manager</h3><p>Another game changer for me was using Zotero. My mentor introduced me to it, and I honestly wish I’d started sooner. Instead of leaving dozens of tabs open, I save papers directly into Zotero.</p><p>A simple colour-tagging system can also make it easier to prioritise your reading. For example, papers marked in red can indicate those that are essential and need to be read as soon as possible, while orange tags can be used for papers that may be relevant and worth exploring later. Green tags can be reserved for papers that are interesting but not immediately important, allowing you to revisit them when time permits. This type of organisation helps prevent reading lists from becoming overwhelming and makes it easier to decide what to read next.</p><p>Without a system, your reading list can become overwhelming very quickly. With a reference manager, everything stays organised and searchable.</p><p>As an added bonus, Zotero makes citation management dramatically easier when writing dissertations, assignments, or journal articles.</p><h3>Build a Daily Reading Habit</h3><p>Another tip is having a dedicated reading time every day.</p><p>I think it’s important to build a small reading habit into your routine. It doesn’t necessarily have to be a full paper every day. Even reading five to ten pages consistently makes a huge difference and by the end of the week, you’ve read lots of pages.</p><p>That’s much easier than trying to force yourself to read seventy pages in a single day over the weekend. Consistency makes managing your reading workload far easier and much less stressful.</p><p>One of the simplest but most effective habits I’ve developed is scheduling a dedicated reading block every day. Consistency almost always beats intensity.</p><h3>Sometimes the Problem Isn’t You</h3><p>One important thing to remember is that sometimes the difficulty lies with the paper, not with you. Academic writing varies considerably in quality. While some researchers are able to explain highly complex concepts in a clear and accessible manner, others can make relatively straightforward ideas seem unnecessarily complicated.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/480/1*VmfKAGzxNVYENqfFrUDynA.gif" /></figure><p>Therefore, if you find yourself struggling to understand a paper, do not immediately assume that you lack the knowledge or ability to grasp the subject. Instead, consider reading another paper on the same topic. You may discover a clearer explanation that makes everything suddenly click. Often, a different author may present the concepts more clearly, providing the explanation you need to fully understand the research.</p><h3>The Best Way to Know If You Truly Understand a Paper</h3><p>Whenever I finish reading an important paper, I ask myself this question, <em>Can I explain this to someone else?</em></p><p>Imagine explaining the study to a friend, colleague, or family member who knows nothing about the topic. Can you explain what the researchers wanted to know, what they did in their study, what they found as a result, and why those findings matter in the broader context of the field.</p><p>If you can answer those questions clearly and simply, you’ve probably understood the paper. If you can’t, there may still be gaps in your understanding.</p><p>Teaching what you’ve learned to others helps it stick more in your memory.</p><h3>Before You Go</h3><p>Reading research papers is one of the most important skills you’ll develop as a student, researcher, or academic. The goal isn’t to read every word of every paper. The goal is to extract useful knowledge efficiently.</p><p>When you start reading strategically, focus on the information that matters, organise what you learn, and build a consistent reading habit, the literature becomes far less intimidating.</p><p>The next time your reading list feels overwhelming, remember that you don’t need to read everything, you simply need a better system for reading.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ad864295b31f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Ethics in AI; Should Machines Make Decisions?]]></title>
            <link>https://medium.datadriveninvestor.com/ethics-in-ai-should-machines-make-decisions-bf743646db43?source=rss-e96e4d3222c2------2</link>
            <guid isPermaLink="false">https://medium.com/p/bf743646db43</guid>
            <category><![CDATA[ethical-ai]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[ethics]]></category>
            <category><![CDATA[deep-learning]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Eniola Ajala]]></dc:creator>
            <pubDate>Mon, 23 Mar 2026 06:37:26 GMT</pubDate>
            <atom:updated>2026-05-13T01:41:49.747Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/735/1*Z6_CnS6YZeUoOrzg_iIHZw.jpeg" /></figure><p>There is a shift happening around us, and most of the time, we don’t even notice it.</p><blockquote>Decisions are being made.<br>Not always by people.</blockquote><p>A system suggests what you should watch next. Another decides whether your loan application gets approved. Somewhere else an algorithm is helping determine who gets shortlisted for a job. None of these feel dramatic in isolation. But taken together, they point to something bigger.</p><p>We are slowly getting used to machines making decisions for us, or about us. And maybe we haven’t paused long enough to ask the more uncomfortable question, should they?</p><h3>The part we don’t question enough</h3><p>It is easy to talk about what AI <em>can</em> do. Every AI system actually carries a set of assumptions. A human making a bad decision affects a few people. An AI system making a flawed decision can affect millions. What’s harder, and far less discussed, is what AI <em>should</em> be allowed to do.</p><p>Because every AI system, no matter how technical it looks, is built on a set of human choices. Someone decides what data is used. Someone defines what “success” looks like. Someone draws the line between what is acceptable and what is not. Someone defines what outcome the system should optimise for.</p><p>And that last part matters more than we often realise. Because when you optimise for something, you also deprioritise something else.</p><p><em>If a hiring algorithm is optimised for efficiency, what happens to fairness?<br>If a healthcare system is optimised for speed, what happens to individual attention?<br>If a recommendation system is optimised for engagement, what happens to truth?</em></p><p>Those decisions don’t disappear once the system is deployed. They actually scale. So when an AI system makes a decision, it’s not neutral. It’s carrying forward the assumptions, priorities, and blind spots of the people who built it.</p><h4><strong>Human decisions are messy, but they are also flexible.</strong></h4><p>We take context into account without always realising it. Tone, intent, history, cultural nuance; these things shape how we interpret situations, whereas AI struggles here. It sees patterns, not meaning. It can detect correlations, but it doesn’t truly understand context the way humans do. That’s why systems can misinterpret sarcasm, cultural references, or language that carries different meanings in different communities.</p><p>What looks like a clear rule to a machine can be a nuanced situation to a human. And when decisions are made without that nuance, the outcomes can feel unfair even when the system is technically “working as designed.”</p><h3>Fairness is not as simple as it sounds</h3><p>There’s a common belief that machines are more objective than humans. That if we remove human emotion, we get fairer outcomes, but AI doesn’t remove bias. It learns from it.</p><p>If a system is trained on historical data that reflects inequality, it doesn’t question that inequality, it reproduces it. Sometimes it even amplifies it. There have been cases where hiring algorithms favoured certain groups over others, not because they were explicitly told to, but because the data they learned from reflected past hiring patterns.</p><p>The system wasn’t “trying” to be unfair, but it was never neutral to begin with. We often say AI should be fair. But fairness itself is complicated. Different groups can have completely different ideas of what fairness looks like. In some cases, treating everyone the same leads to fairness. In others, it reinforces inequality.</p><h3>Can machines make moral decisions?</h3><p>There is a temptation to believe that with enough data and better models, AI will eventually get everything right. But some problems are not just technical.</p><p>Moral decisions often involve values that don’t have a single correct answer. They involve trade-offs that reasonable people can disagree on. No amount of data can fully resolve that.</p><p>So the idea that machines could one day “perfectly” make moral decisions is, at best, incomplete. They can support those decisions. They can inform them, but they cannot replace the human responsibility behind them.</p><h3>The responsibility behind AI</h3><p>If machines are now part of decision making, how do we ensure those decisions are ethical?</p><p>AI is designed, built, and deployed by people. That means every system carries human choices within it; choices about what matters, what should be prioritised, and what outcomes are acceptable.</p><p>AI ethics, sometimes called responsible or ethical AI is about ensuring that these systems are developed and used in ways that align with human values. It’s not just about the final product, but the entire process behind it: the data used, the models built, and how decisions are applied.</p><p>Over time, organisations, governments, and researchers have tried to define what “ethical AI” should look like. While the wording may differ, most of these frameworks revolve around a few key ideas: fairness, transparency, accountability, trust, and human control.</p><p>One way to understand this comes from philosopher Luciano Floridi, who identified five core principles that help guide ethical AI. At its core, ethical AI is about protecting people.</p><p>The first principle is simple, AI should be used for good. It should improve lives, not just optimise systems. Technology should serve humanity, not the other way around.</p><p>Closely tied to that is the idea of avoiding harm. AI systems can affect privacy, job opportunities, personal freedom, and even safety. So it’s not enough for a system to be efficient; it must also be careful. Harm is not always obvious, and sometimes it appears in subtle ways, like bias or exclusion.</p><p>Another important principle is autonomy. As AI becomes more capable, there is a risk of relying on it too much. Ethical AI reminds us that humans should remain in control. Machines can assist, but they should not take away our ability to make independent decisions.</p><p>Fairness is also central. AI systems must not reinforce inequality or discriminate against certain groups. But fairness is not always straightforward, it depends on context, history, and perspective.</p><p>Finally, there is the need for transparency. If a system makes a decision, we should be able to understand how and why it happened. Without this, it becomes difficult to question outcomes or hold anyone responsible.</p><h3>So… should machines be making these decisions?</h3><p>The honest answer is not a clean yes or no. AI is incredibly useful. In some cases, it can even make better decisions than humans, especially when large amounts of data are involved. It can catch patterns we might miss. It can reduce human error. But that doesn’t mean it should replace human judgment.</p><p>There’s a difference between supporting a decision and owning a decision. AI can assist. It can inform. It can guide. But when the outcome has human consequences, there needs to be a human somewhere in that loop, someone who can question, interpret, and take responsibility.</p><p><em>Because machines don’t take responsibility. People do.</em></p><p>Despite these limitations, we still rely on AI systems. So how do we decide whether to trust them? Interestingly, the way we build trust in machines is similar to how we trust people.</p><p>We look for two things, demonstration and explanation. We trust someone if they can show they are capable. And we trust them more if they can explain their reasoning clearly. The same should apply to AI.</p><p>If a system makes a decision, it should not only perform well but also be able to explain <em>why</em> it made that decision. Without that, we are left with a system that produces answers without understanding.</p><p>This is what people often refer to as the “black box” problem. We see the input and the output, but the process in between is unclear.</p><p>And when decisions affect people’s lives, that lack of clarity becomes a serious issue.</p><h3>Before you go</h3><p>Machines are becoming more capable every day. They are learning, adapting, and taking on roles that once belonged only to humans. But capability is not the same as judgment. AI has the potential to transform society in incredible ways. It can improve healthcare, increase efficiency, and make systems more accessible, but they should not replace human judgment entirely.</p><p>Especially in situations that involve ethical consequences, humans must remain involved. Not just as observers, but as decision makers who can question, interpret, and take responsibility. Because when something goes wrong, it is not the machine that is held accountable, it is us.</p><p>Creating ethical AI is not just the responsibility of developers. It requires collaboration between policymakers, researchers, businesses, and communities. Different perspectives are essential, because no single group can define what is fair or acceptable for everyone.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bf743646db43" width="1" height="1" alt=""><hr><p><a href="https://medium.datadriveninvestor.com/ethics-in-ai-should-machines-make-decisions-bf743646db43">Ethics in AI; Should Machines Make Decisions?</a> was originally published in <a href="https://medium.datadriveninvestor.com/">DataDrivenInvestor</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Build an AI Agent using Azure AI Foundry]]></title>
            <link>https://medium.com/@eniola-ajala/build-an-ai-agent-using-azure-ai-foundry-f84b6febd8eb?source=rss-e96e4d3222c2------2</link>
            <guid isPermaLink="false">https://medium.com/p/f84b6febd8eb</guid>
            <category><![CDATA[eniola]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[ai-agent]]></category>
            <category><![CDATA[azureaifoundry]]></category>
            <category><![CDATA[eniola-ajala]]></category>
            <dc:creator><![CDATA[Eniola Ajala]]></dc:creator>
            <pubDate>Sun, 14 Dec 2025 18:14:08 GMT</pubDate>
            <atom:updated>2025-12-14T18:14:08.470Z</atom:updated>
            <content:encoded><![CDATA[<p>Artificial Intelligence (AI) is no longer limited to answering questions or generating text. A growing shift is happening toward systems that can understand intent, make decisions, and carry out tasks on your behalf. These systems are known as AI agents.</p><p>AI agents are designed to go beyond conversation. An agent is built with a specific responsibility in mind. It understands what it’s meant to handle, knows where to get the right information, and can take action when required.</p><p>Microsoft Azure makes this process surprisingly approachable. With the AI Agents Playground, agents can be created, tested, and refined without writing codes.</p><p>In this article, we’ll walk through building a simple support helpdesk AI agent using Azure’s AI Foundry. The agent will answer common support questions using internal knowledge and escalate unresolved issues by triggering a workflow.</p><h3>Setting up the project</h3><p>In the <a href="http://ai.azure.com/?wt.mc_id=studentamb_230833">Azure AI Foundry portal,</a> create a new project. Azure AI Foundry is used to create a new AI resource. This resource lives inside an AI hub and acts as the foundation for working with models, agents, and playgrounds.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*QsGt-DSnB2l262dV2DPAZA.png" /><figcaption>Setting up a new AI project and resource group.</figcaption></figure><p>During setup, a resource group is created. This group becomes the central place where all components related to the agent are managed. Once the resource is created, Azure takes a few minutes to complete provisioning before the tools become available.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2Nsc8RVfSLf5CfGuq4tgFg.png" /><figcaption>Setting up the playground.</figcaption></figure><h3>Entering the Agents Playground</h3><p>With the resource ready, the next step is opening the AI Agents Playground. This is where agents are created, tested, and refined.</p><p>You can see a list of existing playgrounds, but our focus is on agents, so that’s where we’ll click.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*I5tm-wZMBH4JgamJdLTFeA.png" /><figcaption>AI Agents playground</figcaption></figure><p>Before an agent can be created, a model must be deployed. The model serves as the agent’s brain because an agent can’t exist on its own. It needs a model to think with.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gNLi-4m-Zk4vsDwDAOKvyw.png" /><figcaption>Selecting a model for deployment.</figcaption></figure><p>The model is essentially the agent’s brain, so choosing the right one matters. It affects how well the agent understands requests, makes decisions, and responds in different situations. For this setup, a lightweight but capable option makes sense. GPT-4o mini offers a good balance between reasoning ability and cost, which makes it a practical choice for building and testing agents.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*HEhdO90pUykBTswjYe7Ncw.png" /><figcaption>Model deployment.</figcaption></figure><p>Once the model is deployed, the playground becomes active and ready to work with.</p><h3>Threads and Agent Conversations</h3><p>Every interaction with an agent is stored as a thread. A thread records the full conversation, including messages, metadata, tool calls, token usage, and execution time.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AteBKHj4fFsnzuenNNQnew.png" /><figcaption>Agents Playground.</figcaption></figure><p>Starting a new thread creates a clean slate for testing. Older threads remain available and can be reviewed at any time. This makes it easy to understand how the agent behaves across different scenarios and refine its setup as needed.</p><h3>Creating the Support Agent</h3><p>Creating a new agent begins with defining its identity. This includes a name, a description, and a set of instructions.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FTDKWV4RFc-QoLMkjpIHcg.png" /><figcaption>Agent Description.</figcaption></figure><p>For this example, the agent is defined as a first-line support assistant. Its role is to respond to common support questions and escalate issues when they cannot be resolved automatically.</p><p>The instructions explain how the agent should behave, what kind of questions it should answer, and when it should stop responding and escalate instead.</p><h3>Adding Actions</h3><p>Answering questions is only part of what a support agent does. At some point, a request needs to be handed over to a human</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gV-A_AccdivjgP12W4Js7g.png" /><figcaption>Adding action to the agent.</figcaption></figure><p>This is where actions come in. Actions allow the agent to interact with external systems. In this case, a Logic App is connected to the agent. The Logic App is responsible for escalating unresolved issues, such as sending an email to the support team or creating a support ticket.</p><p>A brief description is added to the action to make it clear when it should be used. This helps the agent recognise the right moment to escalate, rather than trying to respond when it shouldn’t.</p><h3>Testing the Flow</h3><p>With the action in place, the agent is tested using realistic support requests. For example, a question that exists in the agent’s knowledge can be answered immediately. A more complex request that cannot be resolved triggers the escalation workflow.</p><p>Looking through the thread logs makes it easy to see how each request was handled; whether the agent responded on its own or called the Logic App.</p><h3>Adding Knowledge</h3><p>To answer questions effectively, the agent needs access to information. This is done by adding knowledge.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kgzPnDFAQPY4VdMyMwxu8A.png" /><figcaption>Adding knowledge to the agent.</figcaption></figure><p>Support documents, FAQs, and troubleshooting guides are uploaded as files and stored in a vector store. These files become searchable knowledge that the agent can reference when responding to support requests.</p><p>The agent is set up to check this knowledge first before replying. If it finds something relevant, it uses it to answer the question. If it doesn’t, the request is escalated instead of guessing.</p><h3>Testing the knowledge</h3><p>After adding knowledge, the agent is tested again. Common questions are answered accurately using the uploaded documents. When a request falls outside the available information, the agent escalates it without hesitation.</p><p>Once the knowledge is added, the agent is tested again. Questions covered by the uploaded documents are answered quickly and accurately. If a request goes beyond what the agent knows, it escalates the issue automatically. This ability to handle routine questions while passing on more complex ones is what makes the agent useful. It takes care of repetitive tasks, while making sure unresolved issues still get the attention they need.</p><h3>Behavior and Performance</h3><p>Every interaction with the agent leaves a clear record. The threads show which knowledge sources were referenced, whether any actions were triggered, and even details like how many tokens were used. Having this visibility makes it easy to understand how the agent is performing and why it made certain decisions.</p><p>Over time, the agent can be improved by refining instructions, adding more knowledge, or adjusting when actions should be triggered to make it smarter and more reliable.</p><h3>Before You Go</h3><p>We’ve barely scratched the surface of what agents are capable of. Agents are enablers, and to learn more, you can explore free resources on <a href="https://learn.microsoft.com/en-us/?wt.mc_id=studentamb_230833">Microsoft Learn</a> or leave a comment suggesting a tutorial you’d like me to write. Keep learning, and cheers!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f84b6febd8eb" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Azure Data Factory: Moving Data from Blob to SQL]]></title>
            <link>https://pub.towardsai.net/azure-data-factory-moving-data-from-blob-to-sql-c460da9c658b?source=rss-e96e4d3222c2------2</link>
            <guid isPermaLink="false">https://medium.com/p/c460da9c658b</guid>
            <category><![CDATA[azure-sql-database]]></category>
            <category><![CDATA[etl]]></category>
            <category><![CDATA[blob-storage]]></category>
            <category><![CDATA[microsoft-azure]]></category>
            <category><![CDATA[azure-data-factory]]></category>
            <dc:creator><![CDATA[Eniola Ajala]]></dc:creator>
            <pubDate>Mon, 03 Nov 2025 16:37:16 GMT</pubDate>
            <atom:updated>2026-05-11T05:07:02.013Z</atom:updated>
            <content:encoded><![CDATA[<p>You know, one thing every data engineer, analyst, or cloud person runs into at some point is moving data. Because let’s be honest, no matter what your project is, data rarely lives in just one place. You’ve got some sitting in an on-premises database that’s been there since forever, a few files hiding out in Azure Blob Storage, and maybe even some bits scattered across other cloud platforms. Somehow, you have to get it from Point A to Point B, cleanly and reliably without losing your mind (or your data).</p><p>And that is why, in this article, I’ll be writing about Azure Data Factory, or ADF as most of us call it.</p><h3>So, What Exactly Is Azure Data Factory?</h3><p>Think of Azure Data Factory like your data delivery manager in the cloud.<br>ADF is the one making sure your data gets where it needs to go, safely, on time.</p><p>What it really does is help you connect, transform, and move data between different places. Maybe you’ve got some data sitting in an SQL Server on-premises, and you need it in Azure. Or maybe you’re moving files from Blob Storage into an Azure SQL Database. Either way, Data Factory handles all that behind the scenes.</p><p>You should know that ADF isn’t where your data lives. It doesn’t store anything. What it does is orchestrate, that is<strong> </strong>it makes sure the data moves smoothly, from one system to another without falling apart halfway.</p><h3>A Simple Analogy:</h3><p>Alright, imagine this.<br>You’re craving a burger from your favorite spot across town. You make an order, and now someone has to get that burger from the restaurant to your doorstep.</p><p>Here’s what’s involved:</p><ul><li>The restaurant’s address and access (so your order can be picked up)</li><li>The actual burger, the thing you want moved</li><li>Your home address (where it’s going)</li><li>And of course, a delivery rider and someone managing the whole process to make sure it all goes smoothly</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*IcpZBrnYD7kaLn6ntcWJBA.png" /></figure><p>So in this illustration, ADF is the manager, it’s just making sure the burger moves from the restaurant to your home safely.</p><h3>Moving Data from Blob to SQL</h3><p>Let’s say, you’ve got a growing business, and every day hundreds of customer orders come in. To keep things simple, your application saves these orders as a CSV file in Azure Blob Storage.</p><p>Now, you want to analyze this data, maybe to find your best-selling products, track returning customers, or calculate weekly revenue. But for that, you’ll need it in a more structured environment, like Azure SQL Database.</p><p>ADF is<strong> </strong>going to be the invisible hand that picks up your “Customer Orders” data from Blob Storage and drops it neatly into SQL so it is ready for analysis, dashboards, and reports.</p><h3>Let’s Get Practical</h3><p>Now that you’ve got a sense of what Azure Data Factory does, it’s time to do some hands-on</p><h4>Step 1: Setting Up the Resources</h4><p>Before ADF can do its magic, you’ll need three things in Azure:</p><ol><li><strong>Azure Data Factory</strong> — the orchestrator that controls the flow of data.</li><li><strong>Azure Blob Storage</strong> — where your source data (the CSV file) lives.</li><li><strong>Azure SQL Database</strong> — the destination where your data will end up.</li></ol><p>First, head to the <a href="https://portal.azure.com/?wt.mc_id=studentamb_230833">Azure Portal</a> and search for Data Factory.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*VRAUpo4moIgh94gzW0QbBA.png" /></figure><p>Click Create, give it a name, maybe something like ADF-CustomerOrders-Demo, and place it in your resource group.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Sg7D4S-YoahWIR-0obYaQw.png" /></figure><p>Choose a region close to you (for example, <em>North Europe</em> if you’re in the UK).<br> Then hit Create,<strong> </strong>it’ll take a minute or two to deploy.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Y5h98vrwSRWWFFIodb2j7w.png" /></figure><p>While that’s loading, go ahead and create two more resources you’ll need:</p><ul><li><strong>Azure Blob Storage</strong> — where your customer_orders.csv file lives, I’ll be using the same resource group, give a name and you can leave the rest as default</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uLbgy7sor3B132NF5-Zo5g.png" /></figure><ul><li><strong>Azure SQL Database</strong> — where the data will end up. Give your database a name, and if you don’t already have a server, go ahead and create one. Then, choose an authentication method that works for you.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kCQuGpbtG3GpXlT3ZMnugA.png" /></figure><p>For the database, choose a simple setup — maybe Standard S1, and remember your username and password (you’ll need them later to connect).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1SD0pdbvvbYsSlpwcnQ_0g.png" /></figure><h4>Step 2: Upload Your Data to Blob Storage</h4><p>After setting up your storage account, open it and navigate to the Containers tab from the menu on the left.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*w1CYock31UtWwv2nD-cxGQ.png" /></figure><p>Create a new container and name it with whatever you like, for this tutorial, let’s call it input, then click Create.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AsTUYc4-hJag0n616Bz0cA.png" /></figure><p>Next, open the input container and upload your data file. I’ll be uploading the customer_orders.csv file. With that, your Blob Storage is ready.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*yM6ajsMikeA6mRJxF4r09w.png" /></figure><p>Click on the file you just uploaded, then select Edit to view it.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Yd0RchKDL_6YdUdBUNNeZw.png" /></figure><h4>Step 3: Prepare the Destination in SQL Database</h4><p>Before we start moving data, we need somewhere to put it.</p><p>Go to your SQL Database in Azure Portal, in the left nav bar, click and open open the Query Editor,</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rZcq4skHwV9D3gHVVszTrw.png" /></figure><p>login and run a quick SQL script to create a table:</p><pre>CREATE TABLE CustomerOrders (<br>    OrderID INT,<br>    CustomerName NVARCHAR(100),<br>    Email NVARCHAR(100),<br>    Product NVARCHAR(100),<br>    Quantity INT,<br>    Price DECIMAL(10,2)<br>);</pre><p>This will be the “home” where your data will land.</p><p>Now we’ve created our data, and we can actually go to Data Factory to start working.</p><h4>Step 4: Connect Blob Storage to ADF</h4><p>Now the fun part begins.<br>Go back to Azure Data Factory and launch studio,</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*99VW2fe-SVzWuTjMl192Lw.png" /></figure><p>this will open the Author &amp; Monitor interface. Here, there are a lot of good stuff.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*JU5oyF7vzPq4kmZ7e9O0nA.png" /></figure><p>Click the pencil icon on the left navigation bar, then take a moment to look around and explore.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6GN4RXOYzIOEJXj95BWoxQ.png" /></figure><p>First, we’re going to create a pipeline, and then we’ll set up the integration runtime.</p><ul><li>In the left panel, go to Manage &gt; Linked Services, then click + New to create your first connection.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*O3LcYUueeYPL6xtj6V38VA.png" /></figure><ul><li>Choose Azure Blob Storage from the list.</li><li>Give it a name (e.g. InputBlob), and under Authentication method, choose Account Key for simplicity.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lWVpXfsIA8dcVV-9zZkyNg.png" /></figure><ul><li>Select your subscription and storage account, then click Test Connection , if it succeeds, hit Finish.</li></ul><p>This is your Linked Service , in the burger story, this is like having the restaurant’s address and access.</p><h4>Step 5: Create a Dataset for the Source Data</h4><p>Next, you’ll define what you’re moving i.e. your “burger.”</p><p>Go to Author &gt; Datasets, click + New Dataset, and choose Azure Blob Storage &gt; CSV.DelimitedText1</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*i2upyFgLyVe-sZ_hN7EKXA.png" /></figure><p>Name it CustomerOrders_CSV, and link it to the InputBlob service you just created.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*J3ffPWYIga2SQRaLx79oEw.png" /></figure><p>Browse to your input container and select the file customer_orders.csv.</p><p>Notice a new dataset was created</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*a4BQRQcSFYJASRAEuu5bdg.png" /></figure><p>Check First row as header, then click OK.<br>You can even preview the data to make sure it looks right, you should see columns like CustomerName, Product, Price, and so on.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*afLBq_Av4CUlrdNBqsYRqA.png" /></figure><h4><strong>Step 6: Create a Linked Service for SQL Database</strong></h4><p>Just like we did for Blob Storage, go back to Linked Services and click + New again.<br>This time, pick Azure SQL Database.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LkapfxliITFpWmn1he_4WA.png" /></figure><p>Name it something like OutputSQL, and under Authentication Type, choose SQL Authentication.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FzSJ3TjuTf3HHaS8O4OFbg.png" /></figure><p>Enter the server name, database name, username, and password that you created earlier.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*y1N8pIuKre9qPz3AmsUuWA.png" /></figure><p>Test the connection, if it’s successful, hit Finish.</p><p><em>This is your destination, the customer’s address where your “burger” (data) will be delivered.</em></p><h4>Step 7: Create a Dataset for the Destination Table</h4><p>Go to Datasets, create a new one, and choose Azure SQL Database.<br>Name it (e.g CustomerOrders_SQL).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AmquLY9Aa9Vznrq7b8ejlw.png" /></figure><p>Under Linked Service, pick OutputSQL.<br>Then select the table you created earlier, CustomerOrders.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*nHD96idmJMAv-8O5HHBfVA.png" /></figure><p>now we have two datasets</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*XHrfzgPzEp_ySicZSQWxQA.png" /></figure><p><em>Now, both sides are connected, we’ve got our source (Blob) and destination (SQL)</em></p><h4>Step 8: Build the Pipeline</h4><p>Head to Author &gt; Pipelines and click + New Pipeline.<br>Give it a name CustomerOrders_ETL.</p><p>From the Activities panel, drag the Copy Data activity onto the canvas.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uZkbv6WT9d_UpdTiZQucVg.png" /></figure><p>In the Source tab, select the dataset CustomerOrders_CSV.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ZdtpVk8PuBNf6bmJjFmL0g.png" /></figure><p>In the Sink tab, choose CustomerOrders_SQL.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*WRadqX1eEVkoFC7NWPFX_g.png" /></figure><p>That’s it, you’ve told Data Factory what to move and where to move it.<br>It’s like assigning a rider (Integration Runtime) to pick up the burger (data) from the restaurant (Blob Storage) and deliver it to your home (SQL Database).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5t_UXCbBhxc8qWT1sChv4w.png" /></figure><h4>Step 9: Debug, Publish, and Run</h4><p>Before publishing, click Debug to test your setup.<br>If everything is configured properly, ADF will start moving your data, you’ll see progress logs in real time.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zMqX8CjY2U7QvY4CkLfHJg.png" /></figure><p>Once it finishes successfully, you can Publish All and then Trigger Now to run the pipeline whenever you want.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wQLxmCqZvqXqOZxSSKYO7A.png" /></figure><p>You can check the Monitor tab to see the status — successful runs will be marked in green.</p><h4>Step 10: Verify the Data in SQL</h4><p>Head back to your SQL Database, open Query Editor, and run:</p><pre>SELECT TOP 10 * FROM CustomerOrders;</pre><p>You should see your customer order data now sitting neatly in your SQL table.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zMqX8CjY2U7QvY4CkLfHJg.png" /></figure><p>Voila you did it, you’ve just completed your first end-to-end data movement with Azure Data Factory.</p><h3>Keeping the Delivery on Track</h3><p>Once your pipeline is ready, you don’t want to manually run it every time new customer order data arrives. Instead, you want it automated.<br>In Azure Data Factory, you can create a trigger. You can set it to run daily, weekly, or even every few hours depending on how often new data lands in your Blob Storage.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_-lPeGWwTaVWcbZEpKhZPA.png" /></figure><p>Now that your deliveries (data movements) are happening automatically, you’ll want to make sure everything’s running smoothly, by making use of monitoring functionality.<br>From the Monitor tab in ADF, you can see a list of all your pipeline runs, their statuses, and the time they took. A green check means your burger (data) was delivered successfully. A red cross means something went wrong, maybe the restaurant was closed (source not available) or the delivery route failed (a step in your pipeline broke).</p><h3>Before You Go</h3><p>This is just the start. There’s actually a lot more you can do with ADF.</p><p>You can learn more about Azure by visiting <a href="https://learn.microsoft.com/?wt.mc_id=studentamb_230833">Microsoft Learn</a>.</p><p>Cheers :)</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c460da9c658b" width="1" height="1" alt=""><hr><p><a href="https://pub.towardsai.net/azure-data-factory-moving-data-from-blob-to-sql-c460da9c658b">Azure Data Factory: Moving Data from Blob to SQL</a> was originally published in <a href="https://pub.towardsai.net/">Towards AI</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to Deploy your Website on Azure App Service]]></title>
            <link>https://medium.com/@eniola-ajala/how-to-deploy-your-website-on-azure-app-service-44d5f18cd842?source=rss-e96e4d3222c2------2</link>
            <guid isPermaLink="false">https://medium.com/p/44d5f18cd842</guid>
            <category><![CDATA[axure-tutorial]]></category>
            <category><![CDATA[azure-app-service]]></category>
            <category><![CDATA[kudu]]></category>
            <category><![CDATA[azure]]></category>
            <category><![CDATA[free-hosting]]></category>
            <dc:creator><![CDATA[Eniola Ajala]]></dc:creator>
            <pubDate>Mon, 27 Oct 2025 09:37:26 GMT</pubDate>
            <atom:updated>2025-10-27T09:37:26.274Z</atom:updated>
            <content:encoded><![CDATA[<p>In my previous article, I wrote about how to host your personal website using Azure Static Web Apps,<strong> </strong>you can read it <a href="https://medium.com/@eniola-ajala/how-to-host-your-personal-website-for-free-on-azure-78938738168e">here</a>.</p><p>While Static Web Apps are great for simple websites like portfolios or blogs, they are not built for more complex or “heavy” sites that need server-side processing, databases, or user logins. on the other hand, Azure App Service is a more robust and flexible way to host your website or web app.</p><p>In this article, I will take you through a more robust way of hosting your website on Azure. We will also consider what happens behind the scenes when you publish an app, clear up the common confusion between App Service and App Service Plan, and finally, I will show you how to host your own app step by step.</p><h3>What is Azure App Service?</h3><p>It’s simply Microsoft’s Platform-as-a-Service (PaaS) offering for hosting web applications, REST APIs, and mobile backends. It allows you to deploy your apps without worrying about managing virtual machines (VMs), networking, and the likes. Since it’s a PaaS offering, Azure App Service takes care of all the underlying infrastructure and maintenance so you can focus on code and configuration.</p><h4>Key Benefits of Azure App Service</h4><ul><li><strong>Quick to set up and scale:</strong> You can deploy your app fast and easily scale it up as your traffic increases.</li><li><strong>Automatic load balancing:</strong> Azure helps manage traffic so your website works perfectly, even with many visitors.</li><li><strong>Works well with GitHub and DevOps:</strong> You can connect it to your code repositories for automatic updates whenever you make changes.</li><li><strong>Custom domains and SSL:</strong> Use your own website name (like <em>mywebsite.com</em>) and secure it with free HTTPS certificates.</li></ul><h3>What happens behind the scenes?</h3><p>When you publish your app to Azure App Service, Azure automatically organizes and runs it in a secure environment. Your app files are stored in the wwwroot folder; this is where all your main files live, such as HTML, JavaScript, and other assets your website needs to work. The App Service runtime host acts as the engine that keeps your app running by receiving requests from users, starting worker processes, and directing traffic to your application.</p><p>Behind the scenes, Kudu (Advanced Tools) helps you manage and troubleshoot your app; you can use it to view logs, check files, run commands, or fix deployment issues. App Settings and Connection Strings store your app’s environment details, such as API keys or database connections, which Azure automatically applies when your app starts running. Finally, the Scaling and Load Balancer features make sure your app performs well by adding more computing power when traffic increases and distributing requests evenly across servers.</p><h3>App Service and App Service Plan, what’s the difference…</h3><p>Think of an App Service Plan as land. Simply, it defines the structure, including CPU, memory, region, and pricing tier. An App Service is like a building; it’s your actual application that sits on that land.</p><p>In other words, your App Service Plan provides the compute resources, and your App Service uses those resources to run your app. You can even host multiple apps under a single plan if they share similar configurations.</p><h3>Let’s Practice</h3><p>I hope you’ve understood the concepts we’ve covered so far. Now it’s time to see how we can actually deploy a web app using Azure App Service.</p><p>We will go step by step, so grab your laptop, open the Azure Portal, and follow along.</p><h4>Step 1: Log in to Azure Portal</h4><p>Go to the <a href="https://portal.azure.com/?wt.mc_id=studentamb_230833">Azure portal</a> and log in to your account.<br>You will see the dashboard, from here, you can search for any service using the global search bar.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*OhVww4Yb4GayeZUe.png" /></figure><h4>Step 2: Create a Resource Group</h4><p>Before creating any resource in Azure, you need a Resource Group. To create one:</p><ol><li>Search for Resource Groups in the search bar.</li><li>Click Create.</li><li>Select your Subscription and give your group a name (e.g., mysite-rg).</li><li>Choose a Region closest to you and click Review + Create.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*H4hfnlTERdG5a3hMI9ElgA.png" /></figure><p>Once validation is complete, click Create again. Your resource group is now ready.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*63Jno9J09zjoz5trZGghjA.png" /></figure><h4>Step 3: Create an App Service Plan</h4><p>Next, let’s create the App Service Plan (the land). You remember it’s the structure that defines your compute resources.</p><ol><li>Search for App Service Plan → Click Create.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2i1IpIwdL6a_Ph6cJtbNsw.png" /></figure><p>2. Choose the Subscription and Resource Group you created earlier.</p><p>3. Enter a name for your website.</p><p>4. Choose the <strong>Operating System</strong> (Linux or Windows).</p><p>5. Select your <strong>Region</strong> (select one closest to you).</p><p>6. Under <strong>Pricing Plan</strong>, choose the <strong>Free tier (F1). </strong>If you have a paid subscription or need more resources, you can choose a higher tier that suits your requirements.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gyadRSnlWjaMwOuflVloxg.png" /></figure><p>7. Click <strong>Review + Create</strong> and wait for deployment to finish.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LQZ4QOciBjhVWe5wF1uTEw.png" /></figure><p>Once the plan is created, you will see the details on your dashboard.</p><h4>Step 4: Create the Web App (App Service)</h4><p>Now it’s time to create your actual <strong>Web App</strong>.</p><ol><li>Search for App Services → Click Create, you will see some options to create but we will select web app for the purpose of this tutorial</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qWOWHRbBr4zqTrivpk4uCg.png" /></figure><p>2. Fill out the following:</p><ul><li><strong>Subscription:</strong> Select your active one.</li><li><strong>Resource Group:</strong> Choose the one you just created (mysite-rg).</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*68f4W-qwPGqe2pFhGwtirA.png" /></figure><ul><li><strong>Name:</strong> Give your app a unique name.</li><li><strong>Publish:</strong> Code.</li><li><strong>Runtime Stack:</strong> Choose the language your app uses (e.g., Node.js, Python, .NET).</li><li><strong>Operating System:</strong> Select Windows or Linux.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3bah70RYx8ZXhg51aWu_zg.png" /></figure><ul><li><strong>Region:</strong> Same as your plan.</li><li><strong>App Service Plan:</strong> Select demo-plan.</li></ul><p>3. Click Review + Create → Create.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gx8jplm6aT8GYVQX1iKSAw.png" /></figure><p>Deployment will start, and once complete, you will be able to view your app’s overview page.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6VFx-mDSqsj1qkfJb68Xzg.png" /></figure><h4>Step 5: Upload Files Using Advanced Tools (Kudu)</h4><p>After your app is created, let’s deploy the website manually using Kudu.</p><ol><li>In the App Service dashboard, in the left nav bar, search for Advanced Tools → Click Go.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*TXPUM5EuHH0-fktV7Ms6Ng.png" /></figure><p>2. A new browser tab opens. in the debug console tab Click on CMD (Command Shell).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Q5rWVqxlah8J-602aS_duA.png" /></figure><p>3. Navigate to the directory:</p><p>site &gt; wwwroot</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*24N1gLxzOpqrPps9uSr3_w.png" /></figure><p>click on site and then wwwroot</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2UFSMATBudWstyWuTaQMQg.png" /></figure><p>4. You will see a default hostingstart.html file. Delete it.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kAD8rQT9Qp1c88AgNl7_Wg.png" /></figure><p>Delete hostingstart.html</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ISoC5OvS3DG215A1YhHodg.png" /></figure><p>5. Drag and drop your web app files (HTML, CSS, JS, etc.) into the folder. <strong>Note:</strong> For this tutorial, I have created a simple demo website. I’ll paste the code files below so you can follow along</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Ru1Lo1NknPGsqoNIDrU2hQ.png" /></figure><p>6. Once uploaded, return to the Overview tab.</p><p>7. Click the Default domain URL to open your app and then you will see your deployed site live on the internet.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*V5wY2rbWj90WfhMeuBJ5oQ.png" /></figure><p>Simple, yeah!</p><h3>Before You Go</h3><p>Using kudu is one of the ways you can deploy your website to Azure App Service, but there are several other ways, and they are all quite easy to implement. For example, you can use ZIP deployment, which lets you package your app into a .zip file and deploy it quickly through the Azure Portal or CLI. Another option is Local Git, where you push your code to an Azure-hosted Git repository, and every push automatically updates your app. If you want me to cover any other method, kindly drop a comment.</p><p>Well, that’s it. Simple, right? Practice and practice. You can learn more about Azure by visiting <a href="https://learn.microsoft.com/en-us/?wt.mc_id=studentamb_230833">Microsoft Learn</a>. Also, follow and watch my page for weekly tutorials on Azure.</p><p>Cheers :)</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=44d5f18cd842" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to Host Your Personal Website for Free on Azure]]></title>
            <link>https://pub.towardsai.net/how-to-host-your-personal-website-for-free-on-azure-78938738168e?source=rss-e96e4d3222c2------2</link>
            <guid isPermaLink="false">https://medium.com/p/78938738168e</guid>
            <category><![CDATA[microsoft-azure]]></category>
            <category><![CDATA[free-hosting]]></category>
            <category><![CDATA[portfolio-website]]></category>
            <category><![CDATA[azure-static-web-app]]></category>
            <category><![CDATA[azure-hosting]]></category>
            <dc:creator><![CDATA[Eniola Ajala]]></dc:creator>
            <pubDate>Wed, 22 Oct 2025 09:48:29 GMT</pubDate>
            <atom:updated>2026-05-11T05:06:46.072Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/736/1*C7h-yieVRv_q50Uvv7WqEw.jpeg" /><figcaption>Microsoft cloud solution provider</figcaption></figure><p>The first time I hosted my portfolio on Azure, I’ll admit, I was a little nervous. The whole process sounded so technical at first, you know… resource groups, repositories, deployments… But to my surprise, it turned out to be so much easier than I expected and when my website finally went live, I just sat there, staring at it with a big smile. It felt amazing to see something I built show up on the internet.</p><p>If you’ve ever dreamed of having your own personal website, maybe a portfolio, blog, or simple landing page, Microsoft Azure gives you a quick and completely free way to make it happen.</p><p>In this article, I’ll walk you through exactly how to host your personal website on Azure, step by step that get you from “I think I can do this” to “Wow, I actually did it.”</p><p>Let’s break it down step by step and see how this works.</p><h3>What Is Azure Static Web Apps?</h3><p>Azure offers a service called <strong>Static Web Apps, </strong>and it is exactly what it sounds like. It’s designed for websites made up of static files like your HTML, CSS, JavaScript. These are called <em>static</em> because they do not need complex server-side processing or databases. There is no backend logic, heavy computation, or APIs running in the background.</p><p>That is why static sites are so easy (and cheap) to host, they are lightweight, efficient, and do not require much computing power.</p><h3>Why the Free Plan Is Perfect for You</h3><p>Azure’s Static Web Apps service has three plans: <strong>Free</strong>, <strong>Standard</strong>, and <strong>Dedicated</strong>. The free plan is perfect for personal projects, portfolios, or small business landing pages.</p><p>Here’s why:</p><ul><li><strong>It is lightweight:</strong> Since your site only contains static files, Azure does not need to allocate expensive resources to run it. That is why they can offer it for free.</li><li><strong>It is reliable:</strong> Azure automatically stores your files and serves them quickly whenever someone visits your site.</li><li><strong>It integrates with GitHub or Azure DevOps:</strong> That means whenever you make a change to your website’s code, your live site updates automatically.</li></ul><p>So, in short, you get free hosting, automatic updates, and enterprise-grade reliability, all in one simple setup. Not bad for a personal website.</p><h3>Understanding the Limitations</h3><p>Of course, the free plan comes with a few limits but nothing that should stop you from hosting. Let’s consider them:</p><ol><li><strong>Limited Bandwidth and Storage:</strong> You get up to 100 GB of bandwidth per month, meaning you will have a monthly cap on how much data your visitors can download and how much space your files can take. But unless you are hosting large videos or getting thousands of daily visitors, you are unlikely to hit that limit.</li><li><strong>Limited Authentication Options:</strong> On the free plan, you can only use prebuilt login methods like Microsoft, Google, or Facebook. You can not create a custom authentication system. For most personal sites, though, you won’t even need login features.</li><li><strong>No Enterprise-Grade Edge Network:</strong> This just means your site won’t have the super-fast, globally distributed performance of the paid tiers. It is still accessible worldwide, it just may not load <em>instantly</em> in every region. This difference is hardly noticeable though.</li><li><strong>No Private Endpoints:</strong> These are advanced security features for corporate or private internal apps. You won’t need them for a personal project.</li></ol><h3>How to Set Up Your Own Static Website on Azure</h3><p>Now that you know what it offers, let’s get your site online :)</p><h4>Step 1: Prepare Your Website Files</h4><p>Make sure you have your HTML, CSS, and/or JavaScript files ready. You can create a simple site with an IDE like Visual Studio Code or even Notepad.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7mKn75IvLKEAkPsC83ncow.png" /><figcaption>Visual Studio Code with website project files</figcaption></figure><h4>Step 2: Create a GitHub Repository</h4><p>Upload your website files to a GitHub repository. Azure will use this to automatically deploy and update your site whenever you push new changes.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jJdKNoEi-6Ui-DVdRdn7QA.png" /><figcaption>Terminal window showing commands for pushing website files from local machine to GitHub repository</figcaption></figure><p>There are a number of ways to push (upload) your files to GitHub, you can do it directly on the website interface, through Visual Studio Code, or by using the terminal.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*M9hVAuAM0FOAdeoY_iyeSQ.png" /><figcaption>GitHub repository showing uploaded files for Azure deployment</figcaption></figure><p>After pushing your files to GitHub, your repository should now contain all your website files just like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*MEe5qHhpqOL4Wk85lMjldA.png" /><figcaption>GitHub repository showing uploaded files for Azure deployment</figcaption></figure><h4>Step 3: Go to the Azure Portal</h4><p>Head to <a href="https://portal.azure.com/?wt.mc_id=studentamb_230833">Azure portal</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*iQ1ZXOgyrgaQqlL-qG2txA.png" /><figcaption>Azure portal dashboard</figcaption></figure><p>On the homepage, click <strong>Create a resource</strong> then search for <strong>Static Web App</strong> and click <strong>Create</strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-ZMQqNUJu3aBTXNyuHTADA.png" /><figcaption>Azure Marketplace page displaying available services, including Static Web Apps.</figcaption></figure><p>Next, select your <strong>Azure subscription</strong>, and if you don’t already have one, create a new <strong>resource group</strong>. Give your website a name for example, <em>demo-portfolio</em> and choose the <strong>Free</strong> hosting plan if you don’t have any subscription.</p><p>Under <strong>Deployment</strong>, select <strong>GitHub</strong> as your source provider, then sign in to connect your repository. Leave the build preset as it is, and when everything looks good, click <strong>Review + Create</strong>, then <strong>Create</strong> to finalize the setup.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LBdJgWPrS7fBWjvUlfJ0cw.png" /><figcaption>Static Web App creation form</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ldCQMx5tiyUsyd85FW0e9g.png" /><figcaption>Static Web App creation form</figcaption></figure><p>Wait a few minutes while Azure sets everything up. Once it is ready, click <strong>Go to resource</strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3QsTidUAk5VMNUy0hIPaDg.png" /><figcaption>Azure deployment completion screen confirming successful creation of the Static Web App resource.</figcaption></figure><p>On the resource page, scroll down to the <strong>Deployment history</strong>. Here, you will be able to see the progress of your site’s deployment and confirm that Azure is building your project from the connected GitHub repository.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*u1MZeC8xU5Ix5-EURjyEXA.png" /><figcaption>Static Web App resource page showing configuration settings, overview, and deployment information</figcaption></figure><h4>Step 4: Watch Your Deployment</h4><p>At <strong>Deployment history tab, </strong>you will see Azure automatically building your website. This is where continuous deployment (CI/CD) happens, it ensures your site updates in real time whenever you make GitHub changes.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*UiXStv8r0V50vFcpp_HWww.png" /><figcaption>Deployment History tab</figcaption></figure><p>Keep refreshing until the deployment process is complete this may take a few minutes. Once it is done, you will see a generated URL for your live website.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dKlZgan5S2KOed6BPBXFzA.png" /><figcaption>Azure Static Web App resource page displaying deployment history and build progress</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ucadoOPil0mLvlMa6rY0IA.png" /><figcaption>live website successfully hosted through Azure Static Web Apps</figcaption></figure><p>Copy the URL and paste it into your browser to view your site in action.</p><p>Congratulations your website is now live :)</p><p>You can now share this link with anyone, and they will be able to access your website from anywhere.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*GK6JnKEap9935cFwPRzEmQ.png" /><figcaption>Azure Static Web App resource page showing deployment details, configuration settings, and overview panel</figcaption></figure><p>And there it is, our portfolio live on the browser. Pretty easy, right?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FzEXBmvQDKlcK-Ti2Cu9qw.png" /><figcaption>Portfolio website successfully hosted and displayed in browser</figcaption></figure><h4>Step 5: Add a Free Custom Domain</h4><p>You can also get a free custom domain on Azure, yes. In the left navigation pane, go to <strong>Settings</strong> and click <strong>Custom Domains</strong>. Then, select <strong>Add</strong> to create your own domain.</p><p>Even on the free plan, Azure lets you use up to two custom domains, that’s pretty cool to me.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*oCSsdVt4X4mZkdmGBrzkVQ.png" /><figcaption>Azure Static Web Apps settings page showing option to add custom domain.</figcaption></figure><p>And that’s it. We are done</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/426/1*-p3KmsGb3E1NU80jr8huUw.gif" /><figcaption>GIF showing the live portfolio in a browser after successful Azure deployment</figcaption></figure><h3><strong>Before You Go</strong></h3><p>Azure Static Web Apps is one of the simplest and most affordable ways to host a personal website. If you have been waiting for the right time to create your portfolio, this is your sign to do it. It is easier than you think and Azure’s free plan makes it effortless.</p><p>Were the steps easy for you? You can learn more about Azure by visiting <a href="https://learn.microsoft.com/en-us/?wt.mc_id=studentamb_230833">Microsoft Learn</a>. Also, follow and watch my page for weekly tutorials on Azure.</p><p>Cheers :)</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=78938738168e" width="1" height="1" alt=""><hr><p><a href="https://pub.towardsai.net/how-to-host-your-personal-website-for-free-on-azure-78938738168e">How to Host Your Personal Website for Free on Azure</a> was originally published in <a href="https://pub.towardsai.net/">Towards AI</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Azure is Essential for Anyone Embracing Cloud Computing]]></title>
            <link>https://medium.com/@eniola-ajala/why-azure-is-essential-for-anyone-embracing-cloud-computing-60b7e7a7a69e?source=rss-e96e4d3222c2------2</link>
            <guid isPermaLink="false">https://medium.com/p/60b7e7a7a69e</guid>
            <category><![CDATA[cloud-computing]]></category>
            <category><![CDATA[virtual-machine]]></category>
            <category><![CDATA[azure]]></category>
            <category><![CDATA[microsoft-azure]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[Eniola Ajala]]></dc:creator>
            <pubDate>Wed, 29 Jan 2025 09:23:48 GMT</pubDate>
            <atom:updated>2025-01-29T09:23:48.648Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*XuuDnKLoJTZbpWq4" /></figure><h3>Why Cloud Computing?</h3><p>As you know, the world is rapidly going digital, and everything is moving at a fast pace. In this article, I will introduce you to cloud computing and Microsoft Azure. We will explore how these two are interconnected and break down the key concepts. Make sure you read this to the end. Now, let’s get started with an example.</p><p>Let’s say you own a local coffee shop and decide to create a mobile app to increase customer loyalty. This app could allow customers to order ahead, earn rewards, and get updates on new menu items. At first, you expect only a handful of regulars to use it, so you might think a small server setup will be sufficient. But imagine your app suddenly becomes popular because a food influencer mentions it online. Now, hundreds, maybe even thousands of people start downloading it and placing orders.</p><p>With traditional server setups, you’d likely hit capacity quickly, leading to app slowdowns, crashes, and frustrated customers. Not to mention, the cost to expand your server capacity to keep up with demand would be massive and take time to implement.</p><h3>The Catch: High Costs and Scaling Challenges</h3><p>While owning your own servers sounds promising, there are two significant downsides. First, the initial setup is costly. Servers, storage, networking, and maintenance add up quickly, and for someone just starting out, that investment can be daunting. Second, it’s risky. If your website doesn’t attract as many users as expected, you won’t be able to recoup those costs. On the other hand, if your website does become famous and attracts high traffic, you’d need to buy even more servers to keep up with the demand, resulting in continuous and unpredictable expenses.</p><h3>The Solution: Cloud Computing</h3><p>This is where cloud computing comes in. The simple solution is Cloud Computing. Cloud computing allows your business to meet customer demand without risking high upfront costs.</p><h3>What is Cloud Computing?</h3><p>When people hear cloud computing, they might picture data floating somewhere in the sky. But in reality, the cloud is not a literal cloud ; it’s a network of physical servers housed in secure data centers across different locations.</p><p>Cloud computing is a platform that gives users access to a vast range of computing resources, like storage, servers, databases, and applications, without having to own or manage any physical hardware.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*YPozVwRgjJUgrmF0" /></figure><p>To understand this better, picture a massive data center with rows and rows of servers, all humming away in a secure building/facility. Cloud service providers like Microsoft Azure, Amazon Web Services (AWS), and Google Cloud manage these centers, offering users remote access to their resources.</p><p>They handle the infrastructure, security, and maintenance, so you can focus on what you want to do, whether that’s developing an app, analyzing data, or hosting a website.</p><blockquote>So, the next time you hear about “the cloud”, remember: your data is not floating in the sky.</blockquote><h3>How Cloud Computing Works</h3><p>It’s quite easy, think of it like subscribing to a streaming service like Netflix. You don’t own the movies or shows, but with a subscription, you can access any content you want when you need it. Similarly, with cloud computing, you pay for access to computing resources and can scale up or down based on your needs. I hope you got that!</p><h3>Flexibility with Pay-As-You-Go</h3><p>Cloud platforms allow you to scale up or down seamlessly, depending on demand. If your website suddenly goes viral, you can quickly increase your resources to handle the traffic without purchasing any new hardware. Conversely, if traffic dips, you can scale back and reduce costs, paying only for what you use.</p><p>Most cloud providers offer a flexible, pay-as-you-go model, which is revolutionary for businesses and individuals. Instead of buying expensive hardware and then hoping it meets your needs, you only pay for what you actually use.</p><h3>What Can Cloud Computing Do?</h3><p>Cloud computing makes it easy for you to have access to advanced technology. Here are some of the incredible things made possible through the cloud:</p><p><strong>1. Machine Learning and Data Analysis</strong><br>Cloud platforms provide access to powerful computing resources that make advanced data analysis and machine learning (ML) accessible to everyone. With cloud-based ML, you can analyze huge datasets to gain insights, make predictions, and make smarter decisions without the need for costly hardware or maintenance.</p><p><strong>2. Data Storage and Backup</strong><br>Cloud computing allows you to store, manage, and back up data effortlessly without needing hard drives or physical storage servers. Cloud platforms (e.g. Azure) keep your data safe and accessible, even if your physical device fails.</p><p><strong>3. Streaming Media Content</strong><br>Have you ever wondered how Netflix, Spotify, or YouTube can serve millions of users globally at any given moment? Cloud computing makes this possible! Media files are stored in the cloud, enabling high-quality streaming without lag or buffering, even during peak usage times.</p><p><strong>4. Creating Software Applications</strong><br>Software development is faster and more collaborative, thanks to the cloud. Developers can work on applications in real-time from anywhere in the world, sharing code and updates without stress.</p><p><strong>5. Automation</strong><br>Automation has become a must-have for businesses looking to streamline operations, and the cloud offers robust tools to support it. Cloud-based automation tools save time and improve efficiency.</p><p><strong>6. Hosting Blogs and Applications</strong><br>You don’t need to own or manage any physical infrastructure. Instead, cloud services provide everything you need to launch and scale your online presence. Cloud hosting is secure, flexible, and cost-effective.</p><h3>What is Azure?</h3><p>Now, lets talk about Azure! Azure is Microsoft’s premier cloud computing platform, designed to empower businesses and individuals with a vast array of resources and services.</p><p>From the example above, using Azure, you can create an app infrastructure that adjusts based on demand. If your app starts gaining popularity, Azure’s cloud services automatically scale up to handle the new traffic without a hitch. When the excitement settles down, your app’s resources scale down to save costs, so you’re only paying for what you need when you need it.</p><p>Azure’s scalability means you can start small but have the ability to handle a big demand spike if your app takes off. Another amazing thing is, it has built-in security, data backups, and analytics, you can be rest assured your customer data is secure and you can analyze user behavior to improve their experience.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/735/0*6RoEOBTtUuR_0-PN" /></figure><p>Microsoft Azure offers a comprehensive suite of over 200 services across 18 categories <em>(how awesome!)</em>, making it one of the most versatile cloud platforms available today.</p><h3>A Brief History of Azure</h3><p>Azure launched on February 1, 2010, Azure began its journey as a robust platform for building and hosting applications in the cloud. Initially branded as Windows Azure, it was reimagined as Microsoft Azure in 2014 to better reflect its evolution beyond just Windows-based services. Over the years, Azure has expanded significantly, adapting to the growing demands of developers and businesses.</p><h3>Key Services of Azure</h3><p>Let’s explore some of the key services offered by Azure in different categories.</p><h4>1. Compute Services</h4><p><strong>- Virtual Machines (VMs):</strong> Azure VMs allow you to create and manage virtual servers, which can run Windows or Linux operating systems.</p><p><strong>- Cloud Services:</strong> Azure’s cloud services facilitate the creation of scalable applications. These services provide a platform to host web applications, services, and APIs without worrying about the underlying infrastructure.</p><p><strong>- Azure Service Fabric: </strong>This microservices platform allows you to build, deploy, and manage scalable and reliable applications. Service Fabric makes it easier to develop complex applications with multiple microservices, ensuring high availability and resilience.</p><p><strong>- Azure Functions:</strong> With Azure Functions, you can easily build serverless applications in any programming language of your choice, such as C#, Python, or JavaScript. This means you only write the code for your specific task or function, and Azure handles the rest.</p><p><strong>2. Networking Services:</strong> Azure’s Networking Services provide essential capabilities to enhance connectivity and performance.</p><p><strong>- Azure CDN (Content Delivery Network):</strong> Azure CDN helps deliver web content, such as images, videos, and scripts, with high bandwidth and low latency to users all around the world.</p><p><strong>- ExpressRoute:</strong> This service allows you to create private connections between your on-premises infrastructure and Azure. ExpressRoute offers more reliability, faster speeds, and lower latencies than typical internet connections.</p><p><strong>- Virtual Network: Azure Virtual Network (VNet):</strong> Provides a secure and isolated environment for your Azure resources to communicate. You can define your own IP address space, create subnets, and establish secure connections to your on-premises network.</p><p><strong>- Azure DNS:</strong> This hosting service allows you to manage your domain names with Azure. It provides ultra-low latency and high availability, ensuring that users can access your applications without delay.</p><p><strong>3. Storage Services:</strong> Azure’s Storage Services are designed to handle various types of data efficiently and securely.</p><p><strong>- Disk Storage:</strong> Azure Disk Storage provides persistent storage for your virtual machines. You can choose between Standard and Premium SSDs, ensuring you get the performance you need for your applications, whether they require high throughput or low latency.</p><p><strong>- Blob Storage:</strong> Azure Blob Storage is optimized for storing unstructured data, such as images, videos, and backups. With features like tiered storage and lifecycle management, you can efficiently manage your data and save costs by moving less frequently accessed data to lower-cost tiers.</p><p><strong>- File Storage:</strong> Azure File Storage offers fully managed file shares in the cloud that can be accessed via the SMB (Server Message Block) protocol. This service allows you to easily share files between applications and users.</p><p><strong>- Queue Storage:</strong> Azure Queue Storage is a service for storing large numbers of messages in the cloud. It enables asynchronous communication between application components, ensuring that your applications can handle peak loads and process requests efficiently.</p><h3>Why Use Azure?</h3><p>Azure, Microsoft’s cloud platform, offers an easy, scalable, and cost-effective solution. With its extensive suite of tools and services, Azure allows individuals and businesses to launch websites, store data, and run applications without upfront infrastructure investments. It’s a flexible, pay-as-you-go model, making it an ideal choice for anyone looking to take their ideas online. Plus, Azure’s global network ensures that your website can be accessed swiftly from anywhere in the world, providing an optimal experience for users.</p><h3>Conclusion</h3><p>Microsoft Azure, long may she reign!</p><p>Azure is an awesome platform that allows you to build and scale your applications without the headaches of managing infrastructure.</p><p>So why wait? Visit the <a href="https://azure.microsoft.com/en-us/get-started/azure-portal">Azure Portal</a> and get your hands on the amazing stuff it has to offer and more.</p><p>Yayy! You made it to the end. I hope you had an interesting read. Check out my previous article on how to use <a href="https://www.freecodecamp.org/news/create-an-ml-model-with-azure-machine-learning-designer/?wt.mc_id=studentamb_230833">Azure for machine learning</a>, where I wrote about the steps you need to take to build your model with Azure ML Designer. Let’s embrace the future with Azure, and who knows what amazing things you’ll create next!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=60b7e7a7a69e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Reflecting on My Experience at the DSN AI Bootcamp 2024]]></title>
            <link>https://medium.com/@eniola-ajala/reflecting-on-my-experience-at-the-dsn-ai-bootcamp-2024-67e1229ff366?source=rss-e96e4d3222c2------2</link>
            <guid isPermaLink="false">https://medium.com/p/67e1229ff366</guid>
            <dc:creator><![CDATA[Eniola Ajala]]></dc:creator>
            <pubDate>Mon, 04 Nov 2024 09:32:45 GMT</pubDate>
            <atom:updated>2024-11-04T09:32:45.515Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4m2knlVhlWFA0iDNU6F0Ww.jpeg" /></figure><p><strong>Dear Readers,</strong></p><p>Today, I will be taking you on a journey through my experience at the Data Science Nigeria AI Bootcamp. This was more than just another training, it was an adventure into the heart of AI and innovation. Unlike my usual articles, this piece will be a little more personal, a bit more storytelling. So, grab a cup of coffee, settle in, and enjoy the ride with me.</p><h4><strong>A Glimpse of the DSN Bootcamp</strong></h4><p>The DSN AI bootcamp, an all-expenses-paid learning bootcamp that builds Africa’s capacity in the use of advanced machine learning and deep learning concepts and drives the application of AI for socio-economic development.</p><p>The bootcamp is driven by a broader strategic intent to accelerate Africa’s development through the solution-oriented application of machine learning to solve social and business problems, and to galvanize a data science knowledge revolution improving employability, technological innovations and sustainable socio-economic development. This reinforces the DSN’s vision to train one million talents in the next ten years through world-class knowledge and best practices applications.</p><h4><strong>My Preparation</strong></h4><p>I first learned about DSN last year, thanks to Olufemi Victor and Abayomi Abiodun, who encouraged me to check it out. I’m so grateful to them for introducing me to this incredible organization.</p><p>This year, I had the opportunity to attend the DSN AI Bootcamp in person at the University of Lagos. I had experienced DSN’s magic before as a virtual attendee in 2023, and that first bootcamp was genuinely life-changing. Building on what I learned then has been pivotal in helping me skill up, but there’s nothing quite like the buzz of being there in person.</p><p>When I first saw the announcement for this year’s bootcamp, it was almost by chance. Just scrolling through WhatsApp, I spotted the flyer with a five-day countdown to the end of the hackathon (<em>a project requirement for selection</em>). I was thrilled but torn; I had so many commitments, and planning for this seemed nearly impossible. Still, my excitement won out.</p><p>I registered for the hackathon that same day and started training the model. I trained it as best as I could in the time I had, made my final submission, and then, fingers crossed and hoped for the best.</p><h4>Getting into the Bootcamp</h4><p>When the congratulatory email from DSN hit my inbox, I could hardly contain my excitement. It was a proud moment for me, knowing all that work had paid off. I started preparing for my journey to the University of Lagos.</p><p>Upon arrival, I was welcomed warmly by the bootcamp organizers, who were just as incredible as I’d imagined. DSN had arranged a comfortable lodge, and I was lucky enough to have wonderful roommates, which made my experience even better.</p><h4>The Bootcamp Experience</h4><p>DSN went all out, bringing in seasoned experts and professionals who delivered sessions that were both insightful and engaging. From general sessions to breakout sessions, each part of the bootcamp taught me something new or inspired me to rethink my approach. I left these sessions not only with new skills but also with renewed motivation to push the boundaries of what’s possible.</p><p>One of the highlights was Cultural Day, where I got more taste of Nigeria’s rich diversity. Seeing the vibrant culture from different tribes was really fun. The bootcamp went beyond just learning; it was an experience that made me appreciate the beauty of cultures beyond my own.</p><p>As you might have guessed, the networking was invaluable. I met some of the brightest minds and made connections that I know will stick with me long after. And, let’s not forget an important part of the bootcamp experience; the food! DSN made sure we were well-fed with delicious meals.</p><p>After the bootcamp, a few of us took a trip to the Lagoon Front on UniLag’s beautiful campus. The view was nothing short of beautiful, a serene scene with the lagoon stretching peacefully before us. It was the perfect way to wrap up an unforgettable week. Just soaking in the scenery and reliving the awesome experiences of the bootcamp. Gosh, I loved every moment!</p><h4>My Gratitude</h4><p>None of this would have been possible without DSN, and I’m deeply grateful for the opportunity. Meeting people I’d only connected with online, finally engaging physically, was amazing. I’ve walked away from the bootcamp with new friends, valuable connections, and a broader perspective.</p><p>This bootcamp gave me something that’s often underestimated which is exposure. It opened my eyes to things I didn’t even know were there and showed me what’s possible with the right mindset and community. I’m beyond grateful I got into this bootcamp.</p><p>This experience has ignited a stronger commitment in me to give back to my community. I’m ready to apply what I’ve learned, to share it, and to collaborate with my connections to make a positive impact.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=67e1229ff366" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to Create a ML Model with Azure Machine Learning Designer]]></title>
            <link>https://medium.datadriveninvestor.com/how-to-create-a-ml-model-with-azure-machine-learning-designer-3a5944ea3035?source=rss-e96e4d3222c2------2</link>
            <guid isPermaLink="false">https://medium.com/p/3a5944ea3035</guid>
            <category><![CDATA[low-code]]></category>
            <category><![CDATA[machine-learning-designer]]></category>
            <category><![CDATA[azure]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[azure-machine-learning]]></category>
            <dc:creator><![CDATA[Eniola Ajala]]></dc:creator>
            <pubDate>Tue, 25 Jun 2024 14:45:24 GMT</pubDate>
            <atom:updated>2026-05-13T01:42:01.076Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Originally published at </em><a href="https://www.freecodecamp.org/news/create-an-ml-model-with-azure-machine-learning-designer/"><em>https://www.freecodecamp.org</em></a><em> on June 25, 2024.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*2MjF0lN8dMicMPio.jpeg" /></figure><p>Did you know that you can create machine learning models without writing any code? If you’re here, you’re probably curious about how to achieve this.</p><p>In this article, I will guide you through building a regression model that predicts automobile prices using Azure Machine Learning’s Low-Code/No-Code tools.</p><p>Regression is a supervised machine learning technique used for predicting numerical values. To understand regression better, you can read my previous article <a href="https://medium.com/@ajalaeniola454/supervised-and-unsupervised-learning-in-machine-learning-06c7151a0c2a">here</a>.</p><h3>Prerequisites</h3><p>To fully understand and follow the steps in this tutorial, you need the following:</p><ol><li>An Azure account: You must have an active Azure account.</li><li>An Azure subscription: An active Azure subscription is required.</li></ol><p>If you don’t have an Azure account yet, you can sign up for <a href="https://azure.microsoft.com/en-us/free/students?wt.mc_id=studentamb_230833">Azure for Students</a>, the <a href="https://github.com/edu/students?wt.mc_id=studentamb_230833">GitHub Student Developer Pack</a>, or an <a href="https://azure.microsoft.com/en-us/free?wt.mc_id=studentamb_230833">Azure Free Trial</a>. These options provide various benefits and free credits to get you started.</p><p>Now let’s get into building our model. Follow the steps below to begin! Make sure you read till the end to learn the entire process.</p><h3>How to Set Up Your Azure Machine Learning Workspace</h3><p>To start working with Azure Machine Learning, you first need to create a workspace. A workspace is a centralized place to manage all the resources and experiments in your machine learning projects.</p><h3>Step 1: Create a Resource Group</h3><p>Start by signing in to the <a href="https://azure.microsoft.com/en-us/get-started/azure-portal?wt.mc_id=studentamb_230833">Azure Portal.</a> Click the Create a resource button (the plus icon) on the left-hand navigation bar.</p><p>In the search bar, type “Machine Learning” and select Azure Machine Learning from the list.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*W-24NM3mO5En3vZW.png" /><figcaption><em>Navigating to the Azure Home page and clicking the plus icon to create a new resource for the machine learning project</em></figcaption></figure><h3>Step 2: Create a New Azure Machine Learning Resource</h3><p>Now click Create to begin setting up your workspace. You will need to fill in the necessary details:</p><ul><li><strong>Subscription:</strong> Select your Azure subscription.</li><li><strong>Resource Group:</strong> Either select an existing resource group or create a new one by clicking Create new and providing a name.</li><li><strong>Workspace Name:</strong> Provide a unique name for your workspace.</li><li><strong>Region</strong>: Choose a region that is closest to your location to reduce latency.</li><li><strong>Container Registry:</strong> Choose Create new unless you have an existing container registry you want to use.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*lwTsTtTcowg3X7de.png" /><figcaption><em>Browsing the Azure Marketplace for machine learning services to find the appropriate tool for building and deploying models.</em></figcaption></figure><h3>Step 3: Review and Create</h3><p>Fill in the form with all the required info. Make sure you provide a unique Name and select a region. Then create your new container registry.</p><p>Review your settings to ensure everything is correct. Then click Review + create to validate your configuration. Once the validation is complete, click Create to deploy your workspace. This process may take a few minutes.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*zuM_ZW2w82-WZhef.png" /><figcaption><em>Setting up a new machine learning workspace in Azure with essential configuration details.</em></figcaption></figure><h3>Step 4: Deployment</h3><p>After the deployment is complete, click the Go to resource button to navigate to your new workspace.</p><p>In the workspace overview, click Launch studio to open the Azure Machine Learning Studio, where you will perform all your machine learning tasks.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*dpMpeNtNeZCHTm-P.png" /><figcaption><em>Resource deployment complete in Microsoft Azure Machine Learning Services.</em></figcaption></figure><h3>How to Set Up Compute Resources in Azure Machine Learning Studio</h3><p>Once your workspace deployment is complete, you need to set up the compute resources required to run your machine learning experiments. Follow the steps below to do this.</p><h3>Step 1: Access the Workspace</h3><p>After the deployment is complete, click the Go to resource button. In the workspace overview, click Launch studio to open the Azure Machine Learning Studio.</p><h3>Step 2: Create a Compute Instance</h3><p>In the Azure Machine Learning Studio, navigate to the left-hand menu and click on Compute. Select the Compute instances tab, then click + New to create a new compute instance.</p><p>Fill in the required details:</p><ul><li><strong>Virtual machine size:</strong> Select Standard_DS11_v2 for a balance of performance at minimal cost.</li><li><strong>Compute name:</strong> Enter a unique name for your compute instance.</li></ul><p>Then click Create set up the compute instance. This process might take a few minutes.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*L0We4RKbbFbvUeHq.png" /><figcaption><em>Configuring compute instances in Azure Machine Learning to ensure the required computational power for training machine learning models.</em></figcaption></figure><h3>Step 3: Create a Compute Cluster</h3><p>In addition to a compute instance, you will need a compute cluster for scalable training. Still under the Compute section, select the Compute clusters tab and click + New.</p><p>Fill in the details:</p><ul><li><strong>Cluster name:</strong> Enter a unique name for your compute cluster.</li><li><strong>Virtual machine size:</strong> Choose Standard_DS11_v2.</li><li><strong>Minimum number of nodes:</strong> Set to 0 to save costs when not in use.</li><li><strong>Maximum number of nodes:</strong> Set to 2 for this tutorial.</li></ul><p>Then click Create to set up the compute cluster. Wait for the cluster to be in a running state before proceeding.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Cl9divtXAWYc7c7M.png" /><figcaption><em>Overview of compute clusters in Azure Machine Learning</em></figcaption></figure><h3>How to Create Your Machine Learning Pipeline</h3><p>In this section, we will create a machine learning pipeline using Azure Machine Learning Designer. Pipelines help streamline the process of preparing data, training models, and deploying them.</p><h3>Step 1: Navigate to the Designer</h3><p>In the Azure Machine Learning Studio, go to the left navigation bar and click on Designer. Then click + New pipeline to start creating a new pipeline.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*4uGjIxwxRii08w0B.png" /><figcaption><em>Interface showing the creation of a new pipeline.</em></figcaption></figure><h3>Step 2: Add Your Data</h3><p>You have the option to use external data or prebuilt sample datasets. For this tutorial, we’ll use a prebuilt sample dataset.</p><p>To use prebuilt sample data, click on the Components tab. Use the search bar or manually scroll to find the Automobile price data dataset. Drag and drop the dataset onto the canvas in the Designer.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*rv5Ow9oxjsKRDsOt.png" /><figcaption><em>Interface showing the addition of Automobile price data set.</em></figcaption></figure><h3>Step 3: Explore the Data</h3><p>Once the dataset is on the canvas, click on it to view its details. Navigate to the Visualizations tab to explore data distribution and statistical summaries. Click the Output+logs tab and then Preview data to inspect the first few rows of the dataset. This will help you understand the data&#39;s structure and identify any potential issues.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Qz2nmhp-fqFZt-gJ.png" /><figcaption><em>Visualization of dataset</em></figcaption></figure><h3>Step 4: Prepare the Data</h3><p>To specify the variables we want to work with, we need to filter out unnecessary columns.</p><p>From the Data Transformation section, drag and drop the Select Columns in Dataset module onto the canvas below the dataset. Connect the output node of the dataset to the input node of the Select Columns in Dataset module.</p><p>Open the module’s settings, select the columns you want to include in your model, and exclude irrelevant columns such as normalized-losses. (<em>We are removing this column because it has high percentage of missing data</em>). Click Save and close to apply the changes.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*ZBVvB7TXh9X-YgBz.png" /><figcaption><em>Interface for selecting columns in the dataset.</em></figcaption></figure><p>Next, you will need to clean any missing data. This helps us handle missing values to improve the model’s accuracy.</p><p>To do this, drag and drop the Clean Missing Data module onto the canvas, below the Select Columns in Dataset module. Connect the nodes and configure the settings to remove rows with missing values. Save and close the settings.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*_kDixUd7ZYJ9HmbA.png" /><figcaption><em>Interface for cleaning columns in the dataset.</em></figcaption></figure><h3>Step 5: Normalize the Data</h3><p>To ensure the model performs well, we need to normalize the data. To do this, from the Data Transformation section, drag and drop the Normalize Data module onto the canvas. Connect the output of the Clean Missing Data module to the input of the Normalize Data module.</p><p>You’ll want to configure the module to use the MinMax scaling method and select the columns to normalize, you can type them out manually by separating each column with a comma. Then save and close the settings.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*GSt4c5VbSBge_9GR.png" /><figcaption><em>Interface for normalizing columns in the dataset.</em></figcaption></figure><h3>Step 6: Finalize the Pipeline Setup</h3><p>Once all transformations are in place, review the pipeline to ensure all steps are correctly configured. Click Submit to run the pipeline. You can monitor the pipeline’s progress in the Jobs section.</p><p>NOTE: Each time you set up a module and you click submit, view the details of the pipeline job to keep track of your flows.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*MRGMaqfdKpZdUk3t.png" /><figcaption><em>Monitoring pipeline through the link</em></figcaption></figure><h3>How to Build the Model</h3><p>This is the fun step where we start building our machine learning model by preparing the dataset for training and applying the appropriate algorithms.</p><h3>Step 1: Split the Dataset</h3><p>Drag and drop the Split Data module onto the canvas. Then connect the output of the Normalize Data module to the input of the Split Data module.</p><p>In the settings panel of the Split Data module:</p><ul><li>Set the Fraction of rows in the first output dataset to 0.7. This means 70% of the data will be used for training the model.</li><li>Set the Randomized seed to any number to ensure reproducibility. For this tutorial, use 123.</li></ul><h3>Step 2: Train the Model</h3><p>Now it’s time to add the Train Model block. Drag and drop the Train Model module onto the canvas. Connect the first output of the Split Data module (the 70% training data) to the input of the Train Model module.</p><p>Click on the Train Model module to configure its settings. In the Edit column section, select the target variable, which in this case is price. Save and close the settings.</p><p>Next, you’ll add the Training Algorithm. From the Machine Learning Algorithms section, drag and drop the Linear Regression module onto the canvas. Connect the output of the Linear Regression module to the right input of the Train Model module.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*rWC2idj78ck7nWmW.png" /><figcaption><em>Interface showing connection of modules in pipeline</em></figcaption></figure><h3>Step 3: Evaluate the Model:</h3><p>Drag and drop the Score Model module onto the canvas. Connect the second output of the Split Data module (the 30% testing data) to the left input of the Score Model module. Connect the output of the Train Model module to the right input of the Score Model module. This will apply the trained model to the testing data to evaluate its performance.</p><p>Next, drag and drop the Evaluate Model module onto the canvas. Connect the output of the Score Model module to the input of the Evaluate Model module. This will provide various evaluation metrics to assess the model’s accuracy.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/508/0*RmzE_NU4EghgAjmv.png" /><figcaption><em>Interface showing connection of modules in pipeline</em></figcaption></figure><h3>Step 4: Review and Submit</h3><p>Double-check all connections and settings in your pipeline. Then click Submit to run the model training and evaluation pipeline. You can monitor the job’s progress in the Jobs section, where you can view logs and outputs.</p><h3>How to Evaluate the Model</h3><p>After training our model, it’s important to evaluate its performance to ensure it’s making accurate predictions. We will use the Score Model and Evaluate Model modules for this purpose.</p><h3>Step 1: Evaluate the Model’s Performance</h3><p>Drag and drop the Evaluate Model module onto the canvas. Connect the output of the Score Model module to the input of the Evaluate Model module.</p><p>The Evaluate Model module will generate various evaluation metrics, such as Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and R-squared (R²), which are critical for assessing the accuracy and performance of the regression model.</p><h3>Explore the Evaluation Metrics:</h3><p>Once the pipeline execution is complete, click on the Evaluate Model module to explore the detailed metrics.</p><p>Here are some key metrics to focus on:</p><ul><li><strong>Mean Absolute Error (MAE):</strong> Measures the average magnitude of errors in predictions, without considering their direction. Lower values indicate better accuracy.</li><li><strong>Root Mean Squared Error (RMSE):</strong> Similar to MAE but gives more weight to larger errors. Lower values are better.</li><li><strong>R-squared (R²):</strong> Indicates how well the model’s predictions match the actual data. Values closer to 1 signify a better fit.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*ixq0-oPTqOAt2D-S.png" /><figcaption><em>Visualization of data output from the regression model.</em></figcaption></figure><p>We just built and evaluated our model. Yayy!</p><h3>How to Deploy the Model</h3><p>Now that we have trained and evaluated our regression model, it’s time to deploy it for real-time inference using Azure Machine Learning. Follow these steps to deploy the model and create an inference pipeline.</p><h3>Step 1: Create the Inference Pipeline</h3><p>Navigate to the Designer section in Azure Machine Learning Studio and create a new pipeline for deployment. Select Real-time inference to set up the deployment pipeline automatically.</p><h3>Step 2: Configure the Pipeline</h3><p>After selecting Real-time inference, the pipeline will be initialized with default components. You can modify the pipeline by adding necessary components.</p><p>We have to make some changes like adding Enter data manually, Execute python script modules. Also, in the select column in dataset block, edit and remove the price column. Remove connection between Score model and web service.</p><ul><li>Enter Data Manually: Drag and drop this block onto the canvas. This component allows manual input of data for prediction. Copy and paste the data below into the data field.</li></ul><blockquote>symboling,fuel-type,aspiration,num-of-doors,body-style,drive-wheels,engine-location,wheel-base,length,width,height,curb-weight,engine-type,num-of-cylinders,engine-size,fuel-system,bore,stroke,compression-ratio,horsepower,peak-rpm,city-mpg,highway-mpg,make 3,gas,std,two,convertible,rwd,front,88.6,168.8,64.1,48.8,2548,dohc,four,130,mpfi,3.47,2.68,9,111,5000,21,27,alfa-romero giulia 3,gas,std,two,convertible,rwd,front,88.6,168.8,64.1,48.8,2548,dohc,four,130,mpfi,3.47,2.68,9,111,5000,21,27,alfa-romero stelvio 1,gas,std,two,hatchback,rwd,front,94.5,171.2,65.5,52.4,2823,ohcv,six,152,mpfi,2.68,3.47,9,154,5000,19,26,alfa-romero Quadrifoglio 2,gas,std,four,sedan,fwd,front,99.8,176.6,66.2,54.3,2337,ohc,four,109,mpfi,3.19,3.4,10,102,5500,24,30,audi 100 ls 2,gas,std,four,sedan,4wd,front,99.4,176.6,66.4,54.3,2824,ohc,five,136,mpfi,3.19,3.4,8,115,5500,18,22,audi 100ls 2,gas,std,two,sedan,fwd,front,99.8,177.3,66.3,53.1,2507,ohc,five,136,mpfi,3.19,3.4,8.5,110,5500,19,25,audi fox 1,gas,std,four,sedan,fwd,front,105.8,192.7,71.4,55.7,2844,ohc,five,136,mpfi,3.19,3.4,8.5,110,5500,19,25,audi 100ls 1,gas,std,four,wagon,fwd,front,105.8,192.7,71.4,55.7,2954,ohc,five,136,mpfi,3.19,3.4,8.5,110,5500,19,25,audi 5000 1,gas,turbo,four,sedan,fwd,front,105.8,192.7,71.4,55.9,3086,ohc,five,131,mpfi,3.13,3.4,8.3,140,5500,17,20,audi 4000</blockquote><ul><li>Execute Python Script: Replace the code in this block with:</li></ul><blockquote>import pandas as pd def azureml_main(dataframe1=None, dataframe2=None): scored_results = dataframe1[[“Scored Labels”]] scored_results.rename(columns={“Scored Labels”: “predicted price”}, inplace=True)<br>return scored_results</blockquote><ul><li>Select Columns in Dataset: Edit and remove the price column to ensure it&#39;s not included in the input data for predictions.</li><li>Remove Connection: Disconnect the Score Model module from the web service output to ensure only the Execute Python Script block connects to it.</li></ul><p>See the images below for a visual representation of how the connections are configured.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/0*lYpV0tVRJdyFsSul.png" /><figcaption><em>Interface of modules connection in pipeline</em></figcaption></figure><h3>Step 3: Submit and Deploy the Model</h3><p>Submit the flow and wait for all processes to complete successfully (green check mark). Once your deployment is successful, you can explore the output to verify everything is functioning as expected.</p><p>You can now deploy the model as a real-time web service for predictions.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*gF4xuEcrdd6ot5ef.png" /><figcaption><em>Configuration settings for setting up new pipeline</em></figcaption></figure><p>Once all the processes are successfully completed in Azure Machine Learning Studio, you are ready to explore the output and deploy your trained model as a real-time web service.</p><p>Here’s how you can proceed:</p><ol><li><strong>Review Pipeline Execution:</strong> Ensure that all the steps in your pipeline have completed with a green check mark indicating success.</li><li><strong>View Outputs and Visualizations:</strong> Navigate through the different components of your pipeline to inspect outputs and visualizations generated during data transformation, model training, and evaluation stages. This helps in understanding how each step has contributed to the overall model performance.</li><li><strong>Check Evaluation Metrics:</strong> Examine the evaluation scores generated by the Score Model block to assess the performance of your trained regression model. Common evaluation metrics include Mean Squared Error (MSE), R-squared (R2), and Root Mean Squared Error (RMSE). These metrics provide insights into how well your model predicts automobile prices based on the input features.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*mFBuOmrJv7Zn2vtU.png" /><figcaption><em>Visualization of data output from model evaluation.</em></figcaption></figure><p>Put the following checks in place before starting the deployment:</p><h4>Prepare for Deployment</h4><p>First, ensure that the Execute Python Script block is correctly configured to transform input data and produce predicted prices.</p><p>You’ll also want to verify that the pipeline is set up to handle real-time inference for deploying the model.</p><h4>Submit the Pipeline</h4><p>Click on the submit button to initiate the deployment process. Wait for Azure Machine Learning to complete the deployment and verify that all components have been successfully deployed.</p><h4>Deploy as Web Service</h4><p>Once deployment is confirmed successful, you can deploy the model as a real-time web service. This web service will be hosted on Azure and can be accessed via API endpoints, allowing applications to send data and receive predictions in real-time.</p><h4>Testing the Web Service:</h4><p>Use the Enter Data Manually block to manually input test data or use external systems to send requests to the deployed web service endpoint. You can verify that the web service responds with predicted prices based on the input features.</p><h4>Monitor and Manage:</h4><p>Monitor the performance of your deployed web service using Azure Machine Learning monitoring tools. You can manage the deployment by scaling resources as needed or updating the model with new data or improved versions.</p><p>Deploying your regression model as a real-time web service enables you to use its predictive capabilities in various applications without the need for direct human intervention. It ensures that your model can continuously provide accurate predictions based on real-time data inputs.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*8voaU3qUeuqcqEjR.png" /><figcaption><em>Interface showing predicted price</em></figcaption></figure><h3>Clean Up</h3><p>The web service you created is hosted in an Azure Container Instance. If you don’t intend to experiment with it further, you should delete the endpoint to avoid accruing unnecessary Azure usage.</p><p>Deleting your compute ensures your subscription won’t be charged for compute resources. You will, however, be charged a small amount for data storage as long as the Azure Machine Learning workspace exists in your subscription.</p><p>If you have finished exploring Azure Machine Learning, you can delete the Azure Machine Learning workspace and associated resources.</p><p>To delete your workspace:</p><ol><li>In the <a href="https://portal.azure.com/?azure-portal=true">Azure portal</a>, in the <strong>Resource groups</strong> page, open the resource group you specified when creating your Azure Machine Learning workspace.</li><li>Click <strong>Delete resource group</strong>, type the resource group name to confirm you want to delete it, and select <strong>Delete</strong>.</li></ol><h3>Conclusion</h3><p>Congratulations! You have successfully created and deployed a regression model using Azure Machine Learning Designer without writing a single line of code.</p><p>In this tutorial, you have learned how to clean and preprocess data, build and train a machine learning model, evaluate its performance, and deploy it as a web service. This low-code/no-code approach in Azure Machine Learning Designer makes it accessible for everyone to use the power of machine learning.</p><p>Remember to clean up your resources to avoid unnecessary charges. With these skills, you can now experiment with other datasets and machine learning problems. Azure’s possibilities are endless, and Azure Machine Learning provides a robust platform for your data science projects.</p><p>Till next time, cheers :)</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3a5944ea3035" width="1" height="1" alt=""><hr><p><a href="https://medium.datadriveninvestor.com/how-to-create-a-ml-model-with-azure-machine-learning-designer-3a5944ea3035">How to Create a ML Model with Azure Machine Learning Designer</a> was originally published in <a href="https://medium.datadriveninvestor.com/">DataDrivenInvestor</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Python Basics and Data Types]]></title>
            <link>https://medium.com/@eniola-ajala/python-basics-and-data-types-f5684942c364?source=rss-e96e4d3222c2------2</link>
            <guid isPermaLink="false">https://medium.com/p/f5684942c364</guid>
            <category><![CDATA[lists]]></category>
            <category><![CDATA[data-types-in-python]]></category>
            <category><![CDATA[data-structures]]></category>
            <category><![CDATA[string]]></category>
            <category><![CDATA[python]]></category>
            <dc:creator><![CDATA[Eniola Ajala]]></dc:creator>
            <pubDate>Tue, 28 May 2024 20:07:45 GMT</pubDate>
            <atom:updated>2024-05-28T20:07:45.922Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*61MVdmfIaLwMoA_B" /><figcaption>Photo by <a href="https://unsplash.com/@rubaitulazad?utm_source=medium&amp;utm_medium=referral">Rubaitul Azad</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>It’s time to learn some Python! This article is an introduction to Python basics and data types. We will be going through operators to perform calculations as well as how Python code can be structured to get the desired output. Before we get started, let’s get familiar with variables.</p><h4>Variables</h4><p>A variable can be considered as a box or storage container for data. A variable is basically for storing information while making it easy for us to refer to that information in our code later. For instance, instead of working with the number 3.14, we can assign it to a variable pi, and use it as many times as possible whenever it is required in our code.</p><pre>pi = 3.142<br>print (pi)</pre><p>The equal sign (=) is used to assign values to variables.</p><p>Print is a function that instructs to print what is inside its parenthesis</p><h4>Arithmetic Operators</h4><p>With Python, we can perform some basic arithmetic tasks such as addition, subtraction, multiplication, division, etc. Let’s consider the examples below:</p><pre>x = 5<br>y = 2<br><br>x + y = 7                                    #additon<br>x * y = 10                                   #multiplication</pre><p>To get the power of a number i.e exponentiation</p><pre>x = 4**2 <br>print(x)                                    #output: 16</pre><p><strong>Modulo</strong>: this shows the remainder of a division</p><pre>g = 18 % 7<br>print(g)                                     #output: 4</pre><h4>Comparison Operators</h4><p>These are used to give two values and boolean (true/false) outputs are given. Using the above variable, x = 5, y = 2 lets consider below examples:</p><pre>g = x &gt; y<br>print(g)                  #output = True<br><br>5 == 2                    #output: False<br><br>2 &lt;= 5                    #output: True</pre><h4>Data Types</h4><p>Data structure and data type are mostly used interchangeably in Python. Actually, data structure is a general computer science term.</p><blockquote>According to Wikipedia, data structure is a data organization, management, and storage format that enables efficient access and modification.</blockquote><p>While data structure is a collection of data values, their relationships, and the functions or operations that can be applied to the data, data type is a term specific to a programming language. In a nutshell, it is a concrete implementation of a data structure in a particular programming language (python in this case).</p><p>There are basic data types like int, float, string, etc. in Python. The data type you use determines how you store it, what you can do to it, and also the value you assign to it.</p><blockquote>You can get the data type of an input by using the type() function.</blockquote><h4>Strings</h4><p>Strings are the opposite of numbers. Strings are a sequence of words like name, address, and anything! It is called a String because it’s composed of characters like alphabets, spaces, punctuation marks, and maybe some numbers.</p><p>In Python, Strings are referred to as strand it is defined within single quotes ( ‘ ), double quotes ( “ ), or even triple quotes ( ‘“ ) depending on the syntax you are comfortable with.</p><pre>name = &quot;My name is John&quot;</pre><p>Let’s consider more strings methods!</p><p>.lower(): This method returns a string with all characters in uppercase converted to lowercase.</p><pre>text = &quot;Hello World&quot;<br>print(text.lower())                                # Output: &quot;hello world&quot;</pre><p>.strip(): This can be used to remove characters from the beginning and end of a string. An argument in the form of a string can be passed to the method to specify the set of characters to be stripped. Without specifying an argument to the method, whitespace is automatically removed.</p><pre>text = &quot;   Hello World   &quot;<br>print(text.strip())                                # Output: &quot;Hello World&quot;<br><br>text = &quot;---Hello World---&quot;<br>print(text.strip(&#39;-&#39;))                              # Output: &quot;Hello World&quot;</pre><p>.title: This string method returns the string in the title case. With .title, the first character of each word is capitalized while the rest of the characters stay in lowercase.</p><pre>text = &quot;hello world&quot;<br>print(text.title())                                # Output: &quot;Hello World&quot;</pre><p>.join(): This concatenates a list of strings together to create a new string.</p><pre>words = [&quot;Hello&quot;, &quot;World&quot;]<br>print(&quot; &quot;.join(words))                              # Output: &quot;Hello World&quot;</pre><p>.upper(): This returns the string with all lowercase characters converted to uppercase.</p><pre>text = &quot;Hello World&quot;<br>print(text.upper())                              # Output: &quot;HELLO WORLD&quot;</pre><p>.replace: This method is used to replace the occurrence of the first argument with the second argument within the string.</p><pre>text = &quot;Hello World&quot;<br>print(text.replace(&quot;World&quot;, &quot;Universe&quot;))  # Output: &quot;Hello Universe&quot;</pre><p>.split: This string method splits a string into a list of items.</p><pre>text = &quot;Hello World&quot;<br>print(text.split())            # Output: [&#39;Hello&#39;, &#39;World&#39;]<br><br>text = &quot;apple,banana,cherry&quot;<br>print(text.split(&#39;,&#39;))         # Output: [&#39;apple&#39;, &#39;banana&#39;, &#39;cherry&#39;]</pre><p>.format(): This method formats a string using placeholders. It is very useful for dynamic string creation.</p><pre>text = &quot;Hello, {}. Welcome to {}.&quot;<br>print(text.format(&quot;Alice&quot;, &quot;Wonderland&quot;)) <br><br><br># Output: &quot;Hello, Alice. Welcome to Wonderland.&quot;</pre><h4>Integer</h4><p>These are whole numbers, positive or negative.</p><pre>x = 100</pre><h4>Float</h4><p>Floats are numbers with decimal fractions. They are real numbers, rational or irrational.</p><pre>y = 77.65</pre><h4>Boolean</h4><p>They have true or falsevalues.</p><pre>g = x &gt; y<br>print(g)                    #output = True<br><br>5 == 2                      #output: False<br><br>2 &lt;= 5                      #output: True</pre><h4>List</h4><p>A list is an ordered sequence of data. Lists in Python can contain different types of data types. A list is declared by using square brackets [ ] and separating the items by a comma.</p><pre>B = [&quot;John&quot;, 3.114, &quot;New York&quot;, &quot;Apples&quot;]</pre><p>An amazing thing about lists is that they are mutable. What this means is simply that the items in a list can be changed/replaced.</p><p>You can add values to the end of a list using the .append() method.</p><pre>B = [&quot;John&quot;, 3.114, &quot;New York&quot;, &quot;Apples&quot;]<br>print(B)<br>#output: &quot;John&quot;, 3.114, &quot;New York&quot;, &quot;Apples&quot;<br><br>B.append(&quot;Rose Flower&quot;)<br>print(B)<br>#output: &quot;John&quot;, 3.114, &quot;New York&quot;, &quot;Apples&quot;, Rose Flower</pre><p>The .count() method searches a list for whatever search term it receives as an argument and then returns the number of matching entries.</p><pre>my_list = [&quot;pencil&quot;, &quot;candy&quot;, &quot;pencil&quot;, &quot;notebook&quot;, &quot;eraser&quot;]<br>pen = my_list.count(&quot;pencil&quot;)<br>print(pen)<br><br>#output: 2</pre><p>To determine the length of items in a list, use the len() function</p><pre>my_list = [&quot;pencil&quot;, &quot;candy&quot;, &quot;pencil&quot;, &quot;notebook&quot;, &quot;eraser&quot;, &quot;textbook&quot;]<br>print(my_list) <br><br>#output: 6</pre><p>The .sort() Python list method sorts the content of the list it is called on. Strings will be sorted in alphabetic order and numerical lists in ascending order.</p><pre>anotherList = [5, 3, 4, 2, 1]<br>anotherList.sort()<br>print(anotherList)<br><br>#output: [1, 2, 3, 4, 5]</pre><h4>Tuple</h4><p>A tuple is an unchangeable ordered sequence of elements. A tuple can also contain multiple data types. Unlike Lists, Tuples are immutable i.e. once a tuple is created, it cannot be changed. A tuple is declared with round brackets ().</p><pre>X = (&quot;John&quot;, &quot;textbook&quot;, 27, &quot;New York&quot;</pre><h4>Dictionary</h4><p>A dictionary is an unordered collection of key-value pairs, where each key is unique. Dictionaries are mutable, meaning that they can be changed after creation. They are often used to store data that can be looked up efficiently by key. A dictionary is declared with curly braces {} and key-value pairs are separated by commas. Keys and values are separated by a colon :.</p><pre># Creating a dictionary<br>person = {<br>    &quot;name&quot;: &quot;John&quot;,<br>    &quot;age&quot;: 27,<br>    &quot;city&quot;: &quot;New York&quot;,<br>    &quot;job&quot;: &quot;Developer&quot;<br>}<br>print(person)<br># Output: {&#39;name&#39;: &#39;John&#39;, &#39;age&#39;: 27, &#39;city&#39;: &#39;New York&#39;, &#39;job&#39;: &#39;Developer&#39;}<br><br><br># Accessing values using keys<br>print(person[&quot;name&quot;])  # Output: John<br>print(person[&quot;age&quot;])   # Output: 27</pre><p>Let’s see how we can modify our dictionary</p><pre># Modifying an existing value<br>person[&quot;age&quot;] = 28<br>print(person)<br># Output: {&#39;name&#39;: &#39;John&#39;, &#39;age&#39;: 28, &#39;city&#39;: &#39;New York&#39;, &#39;job&#39;: &#39;Developer&#39;}</pre><h4>Conclusion</h4><p>Yayy! You read till the end. In this article, we covered the basic things you need to know about Python and its data types to help you understand Python better. You can learn better by practicing these codes in your notebook. If you have any questions, don’t hesitate to drop them in the comments box. Stay tuned for more articles. Till next time. Cheers!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f5684942c364" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>