<?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 FABIYI OPEYEMI on Medium]]></title>
        <description><![CDATA[Stories by FABIYI OPEYEMI on Medium]]></description>
        <link>https://medium.com/@fabiyiopeyemi94?source=rss-343ff6d33d3------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*9m9d5c2sqMx9Hxp2ogD6mQ.png</url>
            <title>Stories by FABIYI OPEYEMI on Medium</title>
            <link>https://medium.com/@fabiyiopeyemi94?source=rss-343ff6d33d3------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 00:04:38 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@fabiyiopeyemi94/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[Adoption Lives and Dies by Advocacy. Here’s How to Build Data Champions]]></title>
            <link>https://fabiyiopeyemi94.medium.com/adoption-lives-and-dies-by-advocacy-heres-how-to-build-data-champions-9e319761dbd5?source=rss-343ff6d33d3------2</link>
            <guid isPermaLink="false">https://medium.com/p/9e319761dbd5</guid>
            <dc:creator><![CDATA[FABIYI OPEYEMI]]></dc:creator>
            <pubDate>Tue, 01 Apr 2025 14:23:30 GMT</pubDate>
            <atom:updated>2025-04-01T14:32:40.028Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fUh_302WVngjzxbvHzrWYg.png" /></figure><h3>No champion? No adoption. No impact.</h3><p>A familiar scene plays out in companies across every industry: a data team builds a robust analytics solution, creates beautiful dashboards, and implements sophisticated data models only to find that months later, hardly anyone is using them. Despite technical excellence, these solutions often fail to gain traction, leaving their potential value unrealized. This common struggle highlights data practitioners’ difficulty of driving adoption among intended users.</p><p>After years of implementing data solutions across dozens of organizations, we’ve observed a critical pattern: projects with engaged internal advocates consistently outperform those without them, regardless of technical sophistication; this observed pattern has been so consistent that it might as well be a law: <strong>successful data projects have data champions</strong>. When it comes to data initiatives, technical excellence alone rarely guarantees success. In summary, this singular factor separates successful data initiatives from those that struggle with adoption.</p><h3>Who is a Data Champion and Why do They Matter?</h3><p>A <strong>data champion</strong> is an internal advocate who believes in the value of data, understands its business impact, and actively promotes its adoption. Importantly, these individuals are <strong>not part of the data team,</strong> but operate within business units. Some champions naturally emerge and are easy to spot, while others require proactive identification and engagement strategies. In this article, I’ll share why these champions matter and provide a practical playbook that discusses our experience in identifying, engaging, and empowering these crucial internal advocates within your organization.</p><p>Research underscores the need for strong data advocacy. A Gartner’s 2018 <a href="https://www.gartner.com/en/newsroom/press-releases/2018-12-06-gartner-data-shows-87-percent-of-organizations-have-low-bi-and-analytics-maturity">survey research</a> found that 87% of organizations have low business intelligence and analytics maturity. Despite lower technical barriers, adoption remains a struggle. McKinsey <a href="https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-data-driven-enterprise-of-2025#/">predicted</a> that by 2025, broad organizational data literacy and advocacy — what they call “Culture Catalysts” — will be crucial for widespread analytics adoption.</p><p>While data leaders often emphasize self-service analytics and data-driven decision-making, achieving this vision in reality is more challenging due to the complexities and nuances involved, requiring more than just tools and technology. However, one factor we’ve seen across different organizations where we’ve implemented successful data solutions in their data maturity journey, is identifying these individuals we call “Data Champions,” who help bridge the gap between ambitious goals and reality. Technical excellence matters, but human advocacy often determines whether data solutions drive business value. Yet surprisingly, many Data professionals and Analytics leaders focus exclusively on technical excellence while overlooking the human element that often determines success.</p><h3>Identifying Potential Champions: What to Look For</h3><blockquote><em>The board and the CEO raise the data clarion, and the people on the front lines take up the call. But to really ensure buy-in, someone’s got to lead the charge. That requires people who can bridge both worlds-data analytics/science and on-the-ground operations. And usually, the most effective change agents are not digital natives.<br>Source : </em><a href="https://www.mckinsey.com/capabilities/quantumblack/our-insights/why-data-culture-matters"><em>Mckinsey — Culture Catalyst</em></a></blockquote><p>Not every organization has obvious data champions waiting in the wings. Often, you need to find and nurture these relationships. It’s important to note that not everyone with a leadership title or technical savviness makes an effective data champion. However, the most effective champions tend to share common traits:</p><ol><li><strong>Curious Questioner</strong></li></ol><p>These are individuals who are naturally curious about the numbers and metrics. Not everyone who interacts with reports is interested in the exploration and deep dive to see insights or patterns, or is interested in a spike or downward trend in the data. The curious questioners are the ones asking insightful questions about metrics, even if they lack technical data skills. So you want to look for these individuals who consistently ask “why” questions about business metrics. They might say things like: “<em>Do we know what’s driving this trend</em>?” or “<em>How does this compare to last quarter</em>?” or “<em>How can this number be cross referenced with other operation tasks for visibility?</em>”. These natural analytical thinkers often become your strongest champions and their analytical thinking helps create a fertile ground for data advocacy.<strong>‍<br>‍</strong></p><p><strong>2. Business Pain Awareness</strong></p><p><strong>‍</strong>The most effective champions deeply understand organizational pain points that data could solve. These individuals constantly acknowledge the limitations of the current workflow of their daily operations. Sometimes they express some frustration about lacking visibility into some aspect of their workflow, or how onerous a task it is for them to stitch together their manual reports across multiple sources to do their job. You can sometimes hear them say, “<em>I know we could make better decisions if we have more visibility into this</em>”, or “<em>I’m blind to how this aspect of the business is doing</em>”, or “<em>I wish there’s some automated workflow available to remove this daunting task to help create more leadway and free my team of some time for another productive task</em>”. These are some statements you shouldn’t ignore as you engage with different stakeholders within your organization. These individuals are potential change agents for your analytics solution adoption.<strong>‍<br>‍</strong></p><p><strong>3. Organizational influence</strong></p><p>Champions need influence to be effective. However, this influence doesn’t necessarily mean formal authority — sometimes, a respected mid-level or senior manager has more operational influence than a distant executive. These individuals can influence others formally or informally.</p><p>We’ve worked with a client where the data champion led a Looker training for their team, prepared a training manual/deck, and only asked for our team to join the call to help support and provide responses to questions they were unable to answer if needed. We’ve also worked with another client where the data champion helped drive adoption of the solution we implemented, with different stakeholders actively engaging with the reports. However, we noticed a significant decline in solution usage during a period when our data champion was away for an extended time.</p><p>These scenarios provide insight into how critical having data champions within your organization and across different functions can be. Champions who have the credibility and organizational influence to function as change agents help drive the required adoption. If you are able to identify these individuals, your efforts are minimized as they relate to driving awareness, utilization, and adoption of the great analytics solutions your data team has implemented.<strong>‍<br>‍</strong></p><p><strong>4. Communication Fluency and Growth Mindset</strong></p><p>Great champions translate between technical and business languages. They understand enough about data to appreciate its potential while speaking in terms of outcomes that resonate with decision-makers. In addition, they embrace learning and adapt their understanding as initiatives evolve. They’re comfortable saying “I don’t know, but let’s figure it out” or “Here’s what I’m thinking, but I don’t know how to get that done or if there’s any data available.”Given that data champions are not technical folks and are entrenched more in the business than the data team, they are able to think from a more strategic perspective. Most times, they can provide pointers and directions for new initiatives, reports, and solutions to implement that will drive value organization-wide</p><h3>The Data Champion Identification Checklist</h3><p>Having provided context on the traits to look for in a potential data champion and how to identify a potential change agent within an organization as a data leader, you can use this simple summarized checklist and assessment to identify potential champions in your organization:</p><ul><li>Shows genuine curiosity about metrics and performance trends</li><li>Expresses frustration with current data limitations</li><li>Has influence (formal or informal) with key stakeholders</li><li>Demonstrates comfort discussing both business and technical concepts</li><li>Exhibits willingness to try new approaches</li><li>Occupies a position where data insights could visibly improve outcomes</li></ul><p>You can spot these potential data champions within your organization by starting to look for people who say things like:</p><ul><li>“I wish we had better visibility into…”</li><li>“I’m tired of making decisions based on gut feeling when we should have data.”</li><li>“If we could just see how these metrics connect…”</li><li>“I spend hours manually compiling reports that should be automated.”</li><li>“We’ve tried to be more data-driven before, but we need a better approach.”‍</li></ul><h3>Champions don’t Care about your Tech. They care about their pain.</h3><p>Once you’ve identified potential champions, how do you engage them effectively?</p><h4>Start with Their Pain, Not Your Solution</h4><p>Many data initiatives begin with technical capabilities rather than business needs. Reverse this approach:</p><ol><li><strong>Conduct a “pain interview”</strong> focusing on their biggest information gaps</li><li><strong>Document current manual processes</strong> that could be automated</li><li><strong>Explore decisions they’re making with insufficient information</strong></li><li><strong>Identify metrics they wish they could track</strong> but currently cannot</li></ol><p>Schedule deep-dive conversations to get more insight into areas where the data champions would need analytics support (most times they would have already spoken about this as part of their business pain awareness traits from a data limitation perspective). When working with a Storage Company, we discovered the COO lacked visibility into critical aspects of business operations, which gave my team an area to provide quick wins and show value. Similarly, in the context of the data champion mentioned earlier who took the initiative to organize a Looker training for her team, when we started the engagement she had communicated challenges with manually stitching together multiple reports across different data sources. She also mentioned limitations that came with that approach as they relate to the ability to create more granular reporting. Focusing on the pain of your data champion helps strengthen the relationship, and when they begin to see value, they naturally want to drive the adoption of the solution.</p><h4>Deliver Quick, Impactful Wins</h4><p>Your Champions need evidence to advocate for your initiatives. Create early deliverables that demonstrate value, even if they’re simple:</p><ul><li>A basic dashboard showing previously unavailable insights</li><li>Automation of a manual reporting process</li><li>Visualization of a key metric over time</li></ul><p>It’s important to note that delivering quick and impactful wins may sometimes come at the expense of perfectly built data infrastructure. However, a workable solution that drives business value usually outweighs technical perfection. Sometimes, the “least bad solution” within constraints is the best path forward. While embracing a hacky approach isn’t easy, especially knowing it might create technical debt, speed and flexibility often matter more than perfection when driving immediate business value. These early wins provide your champion with tangible evidence to share with executives, building momentum for broader initiatives. Embrace pragmatism over perfection when delivering quick wins.</p><h4>Communicate Efficiently and Strategically</h4><p>Respect your champion’s time constraints while keeping them informed and engaged:</p><ul><li>Deliver high-value updates in their preferred format (brief emails, one-page summaries, or quick stand-ups)</li><li>Focus communications on business outcomes rather than technical details</li><li>Create simple visual progress trackers they can quickly review</li><li>Prepare “elevator pitch” updates they can easily share with other executives</li></ul><h4>Align with Their Strategic Objectives</h4><p>For senior champions like COOs or Senior Managers, focus on strengthening the connection between data initiatives and their strategic priorities:</p><ul><li>Connect data wins directly to their performance metrics and objectives</li><li>Prepare materials they can use in board meetings or executive presentations</li><li>Identify opportunities where data insights can enhance their strategic decision-making</li></ul><h4>Equip them with Knowledge</h4><p>Help your champion become more data-literate through informal teaching moments. Explain concepts clearly without condescension, and gradually increase complexity as their understanding grows. A champion who can confidently explain the “why” behind a data initiative becomes exponentially more effective. Create glossaries of concepts or summarized documentation to clarify any technical difficulties.</p><h3>The Signs Your Champion Is Making an Impact</h3><p>How do you know if your champion relationship is effective? Look for these indicators:</p><h4>Engagement Indicators:</h4><ul><li>Proactively requests enhancements or new capabilities for existing solutions</li><li>Asks thoughtful questions that demonstrate deepening understanding</li><li>References data insights in communications with their team and leadership</li><li>Volunteers to demonstrate the solution to colleagues or explain its value</li></ul><h4>Adoption Indicators:</h4><ul><li>Increased usage metrics for data tools within their sphere of influence</li><li>Higher attendance and participation in data-related meetings</li><li>Growing number of cross-functional requests for access</li><li>Expanded use cases beyond the initial implementation</li></ul><h4>Impact Indicators:</h4><ul><li>Business decisions explicitly reference data insights from your solutions</li><li>Champion independently identifies new applications for the data</li></ul><p>When these indicators emerge organically, you’ve moved beyond simply having a supporter, you’ve developed a true champion who is actively driving data culture within their organization.</p><h3>Champions as Transformation Catalysts</h3><p>Time and again, we’ve seen that the <strong>right combination of strong data advocacy and technical excellence transforms organizations.</strong> A well-built analytics solution is just a starting point. What truly drives adoption is an internal network of champions who make data part of everyday decision-making.</p><p>By systematically identifying, engaging, and empowering these critical advocates, you dramatically increase the likelihood of your data initiatives delivering real business value. More importantly, you help create a data culture that extends beyond any single project.</p><p>The next time you plan a data initiative, ask yourself: “Who will champion this work?” The answer to that question may determine your success more than any technical decision you make. Data champions are the key to analytics adoption, so don’t just focus on tools, invest in building a network of champions who can drive real impact!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9e319761dbd5" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Getting a Data related role is Hard!!!]]></title>
            <link>https://fabiyiopeyemi94.medium.com/getting-a-data-related-role-is-hard-cc5ed6cbdd85?source=rss-343ff6d33d3------2</link>
            <guid isPermaLink="false">https://medium.com/p/cc5ed6cbdd85</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[data-analyst-jobs]]></category>
            <category><![CDATA[data-science-job]]></category>
            <dc:creator><![CDATA[FABIYI OPEYEMI]]></dc:creator>
            <pubDate>Sat, 23 Jan 2021 18:38:10 GMT</pubDate>
            <atom:updated>2021-01-23T18:38:10.990Z</atom:updated>
            <content:encoded><![CDATA[<p>Be prepared for the reality of getting a Data Analyst/Science Job</p><p>Have you ever asked yourself the question if data science is in demand, then why is it so hard to get a data analytics/science job? When people ask me ‘Ope, can I learn Data Science in 6 months rigorously &amp; get a related role?’, well such questions I really understand are a product of information from an academy or online courses that preaches the good news of how &amp; why having a data skills make you a sought after professional, but the truth is, it’s far from the reality. It’s okay to give a timeline and target but don’t get discouraged when the job doesn’t show up at the targeted time, stay persistent and dogged.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0lT5GPuL7Y2OyK7Z-IOQdg.jpeg" /></figure><p>I attended quite a couple of data science interviews last year(you will be shocked if I tell you the actual number) with a 98% rejection rate, well I know I didn’t share those failures &amp; rejections on LinkedIn even though I shared the good news of the success, this should help you see the reality so you shouldn’t get intimidated with the congratulatory &amp; success feeds or frustrated when you haven’t received any congratulatory email after several trials based on what you see on a social platform like this because majority won’t post the sad moments.</p><p>That said, with my little experience in my current role, I have come to realize that finding a data scientist is hard &amp; hiring one is much harder, and the NEWER you are in the field the harder getting your foot in the door is likely to be. Landing a job for most data science and analyst roles means experiencing a lot of rejection. You must be PERSISTENT.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*N6m-w3MK9btWMciMgZRJyg.jpeg" /><figcaption>Don’t get frustrated or Discouraged</figcaption></figure><p>The reality is that some companies will not respond to your application, or may not even send you any feedback whether a rejection, success or update after completing their on-site interviews, They may just leave you hanging in that state of unknowing and wondering what you did right or wrong in the interview process, constantly rehearsing how your interview went and how you think you should have answered the questions, I still have many of such Interviews I attended and still no response till date.</p><p>Don’t get discouraged or frustrated and lose motivation to stop learning because from experience lethargy will set in and your enthusiasm to continue to learn might drop, but don’t stop, your goal as a candidate is to not give up no matter what the process looks like as I often tell my friends to see rejections as an inspiration and channel the energy to other applications, so you’ve got to develop a shock absorber for rejections. Some people are lucky to land a role quickly, but I can say for a fact with so much gusto that they are the minority (an Outlier using the data lingo).</p><p>This post is to help you prepare your mind for what the reality is about securing that data science or analyst role and If you are in that stage don’t get discouraged, keep trying and never give because you wouldn’t know the opportunity that will meet your preparedness.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Ad_HhE0iXo2YHZW_8G1icg.jpeg" /><figcaption>Don’t give up, Keep Trying!</figcaption></figure><p>In your search for that job, the following can help get you to the interview door because truthfully many organizations find it difficult recruiting the right skilled person for the job because a lot of unskilled folks who claim the title;</p><ul><li>Join a data community and be an active member: This is one of the easiest ways to be abreast of recent data roles in different organizations and I will encourage you to be an active member of the community as it makes recommending you easy for a role when recruiters reach out to the founders of the community for talents.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lLfD7uSdFK_L6eePdcVbaw.jpeg" /><figcaption>AISaturday Lagos Cohort 3</figcaption></figure><ul><li>Pitch your skills on Social Media — Create awareness about your skills on LinkedIn, Twitter, Medium, etc.</li><li>Attend Meetups &amp; Tech events</li><li>Keep learning &amp; don’t just take online courses but work on a project that showcases the lesson learned from the courses.</li></ul><p>The above list isn’t exhaustive but my curated list that was helpful during my job hunting season which helped me maintain enthusiasm.</p><blockquote>Nobody said it would be easy, so don’t ever give up when it gets Hard</blockquote><p>Let’s connect on;</p><p><a href="https://twitter.com/Opiano_1">Twitter</a></p><p><a href="https://www.linkedin.com/in/opeyemifabiyi">LinkedIn</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cc5ed6cbdd85" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Datathon2020 — Predicting weather disruption of public transport]]></title>
            <link>https://fabiyiopeyemi94.medium.com/datathon2020-predicting-weather-disruption-of-public-transport-86373a19e6bc?source=rss-343ff6d33d3------2</link>
            <guid isPermaLink="false">https://medium.com/p/86373a19e6bc</guid>
            <category><![CDATA[exploratory-data-analysis]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[FABIYI OPEYEMI]]></dc:creator>
            <pubDate>Tue, 19 May 2020 00:09:26 GMT</pubDate>
            <atom:updated>2020-05-22T20:02:34.481Z</atom:updated>
            <content:encoded><![CDATA[<h3>Datathon2020 — Predicting weather disruption of public transport</h3><h4>Don’t Just Learn Data Science, Solve Business Problems</h4><p>There is no Data Science if one is not solving a problem because data science is not about the fanciful algorithm or statistical modeling but the end goal is the solution proffered to a problem, and it is no coincidence that Business understanding top the list in the Cross-Industry Standard Process for Data Mining which is the general adopted standard process for solving data-related problems that emphasize the need to focus on solving a problem with data.</p><p>This project was inspired by the just concluded <a href="https://www.datasciencesociety.net/datathon/">Datathon</a>, an International annual 48-hour Data Science Hackathon which brings all Data Scientists from all around the world who are thrilled to be challenged with real data science cases.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5tVh_IIrPSceM4bWQfY7IQ.png" /><figcaption>Datathon 2020</figcaption></figure><p>This year’s Datathon featured four(4) <a href="https://www.datasciencesociety.net/category/datathon-cases/">business cases</a> for participants to select any case of their choosing to leverage on their data skills to proffer solutions of which one of the case studies I selected will be presented in this article on my approach and solution.</p><p><strong>CASE STUDY</strong></p><p>Climate change is projected to increase the frequency and intensity of some extreme weather events which is likely to damage transportation infrastructure and cause a disruption in the public transport and increase the risk of delays and failure due to storm, flooding and higher temperatures affecting the reliability and capacity of the public transportation.</p><p>The Goal</p><p>Can you analyze the weather data to predict public transport service disruption in Dubai? How can we plan for less disruption in the wake of severe weather conditions and leverage the emergency management plan as well as providing uninterrupted services and products to citizens?</p><p>The solution to the case given will follow the CRISP-DM approach which includes;</p><ol><li>Business Understanding</li><li>Data Understanding</li><li>Data Preparation &amp; Analysis</li><li>Data Modelling/Validation</li><li>Result</li><li>Deployment</li></ol><p><strong>Business Understanding</strong></p><p>The goal of the case study is to predict public transport disruption using weather conditions information. A grasp understanding of the problem to be solved drive the whole analysis to be done.</p><p><strong>Data Understanding</strong></p><p>A JSON weather dataset was provided, however, no information on public transport was available which implies that there is a need to source for external data to have adequate information to solve the problem. The Business case was for Dubai provided by Ernst &amp; Young, so I explored the open-source data repository of Dubai to download all the public transports information which covers 2 years. <br>The two (2) data sets — weather &amp; transports were loaded to pandas data frame as shown below</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/d72c7144b6f33fbe6c97e65b68c4d6f1/href">https://medium.com/media/d72c7144b6f33fbe6c97e65b68c4d6f1/href</a></iframe><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4bspC7wFawXPEbfP4bfZHA.png" /><figcaption>Weather &amp; Transport data head</figcaption></figure><p>The weather dataset includes all the Dubai weather condition from 2018 till date and the transport dataset includes all the number of trips completed for all the Dubai public transport type within the same year in view, and as displayed above it is clear that the weather dataset requires some cleaning before any analysis can be performed.</p><p><strong>Data Preparation &amp; Analysis</strong></p><blockquote>It is often said that generally Data Scientists tend to spend 70% of the project time on data preprocessing , exploring the data and various treatment.</blockquote><ul><li>Data Preparation</li></ul><p>The labels of the Main, wind, clouds &amp; weather features are key-value pairs dictionary, so we can use the Pandas series extract function.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/040c3b9f5c8874410c1d8c68e8110c47/href">https://medium.com/media/040c3b9f5c8874410c1d8c68e8110c47/href</a></iframe><figure><img alt="" src="https://cdn-images-1.medium.com/max/828/1*A6Y0AKqspHIYtzcZ0zNAcA.png" /><figcaption>Extracted fields DataFrame</figcaption></figure><p>After extracting the key-value pairs dictionary label and converting the result to a data frame included in the dataset, a further data cleaning still required to make the data fully prepared for any form of analysis and modeling as seen below;</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/74abbdc14c9eede9db3bcb67bf412651/href">https://medium.com/media/74abbdc14c9eede9db3bcb67bf412651/href</a></iframe><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*z3EOAOeZAwi59Kl_vmxwkg.png" /><figcaption>Weather Data</figcaption></figure><p>Our weather dataset is almost ready to be used for analysis and modeling, from domain knowledge acquired while researching during the course of the hackathon, it was discovered that the icon feature had some implication where ’n’ is related to <em>night &amp; ‘d’ day </em>which led to engineering a new feature called <strong>period </strong>as shown below;</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/d71f2c0a60d1d43f17f2165b20c33933/href">https://medium.com/media/d71f2c0a60d1d43f17f2165b20c33933/href</a></iframe><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AC4UUMnlo2SjlTXP5q0YUw.png" /><figcaption>Cleaned Weather data</figcaption></figure><p>After successfully preparing the weather dataset, the public transport data can now be merged to have a complete dataset to use for the analysis using the unique Id created in the transport &amp; weather data and filling the missing information respectively using mean imputation given that the proportion of the missing information is very infinitesimal.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/a7c0dea5278c96b2c88841a79461d27a/href">https://medium.com/media/a7c0dea5278c96b2c88841a79461d27a/href</a></iframe><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*yawecdw8meNnzmhkaEdUeQ.png" /><figcaption>Cleaned Merge Data</figcaption></figure><ul><li>Exploratory Data Analysis (EDA)</li></ul><blockquote>If you torture the data long enough, it will confess to anything</blockquote><blockquote>~ Ronald Coase</blockquote><p>After cleaning the data and preparing it, Exploratory Data Analysis (EDA) was then performed on the data to gather insights that will be useful for the model to learn from the data to help improve the performance of the model.</p><p>The business problem was to predict the disruption of public transport by weather conditions, so the analysis performed was done with a focus on this problem, hence the reason to make <strong>trips completed</strong> in the transport data the target label our Machine learning model will be making predictions on using the explanatory features of the weather conditions.</p><p><em>Univariate Analysis</em></p><p>An analysis of the year was performed to determine which year had the highest number of trips completion across all transport types.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/576/1*_PFK24VsKH7TXCoYbHN-FA.png" /><figcaption>Trips Completed per year</figcaption></figure><p>As seen from the displayed output we see that 2018 had the highest number of trips completed with a decline in 2019 and 2020. It is therefore of importance to perform a diagnostics analysis to understand why the decline in 2019 to help in preventing the disruption in 2020 which is already imminent.</p><p>In answering the questions of how and what happened, we checked ;</p><ul><li>The Weather Condition for High Trips completed for the major transport type in 2018 &amp; 2019</li><li>The weather condition for fewer Trips completed for the major transport type in 2018 &amp; 2019</li><li>Compare the above insights from 2018 &amp; 2019 to check the current status of 2020.</li></ul><p>Insights from the diagnostic analysis will be useful in engineering new features that will be used to train the machine learning model to make predictions of what the transport trips completion status will be with respect a given weather condition</p><p><em>Public Transport Analysis</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/576/1*JFlZwc16PI9WUuVVZ-K3gQ.png" /></figure><p>From above we see that Metro and Bus transport type had the highest trip completion across the years, Marine and Tram had little trip completions, hence we can combine both Marine &amp; Tram to a unique label Others to help improve the machine learning model</p><p>We then checked which period of the day was trip completed most across all the years</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*w2s-QUBZfzEI4yrEU27DBw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*ylxIrcPDuyeptpSGfoNlTQ.png" /></figure><p>As seen from the chart that in 2018 there seems to be an even distribution in the number of trips completed during the day and at Night for both Metro and Bus which are the most patronized transport type. In 2019 more trips were completed during the day, however in 2020 so far trip completion in the night is little more than the day</p><p><em>Weather Conditions Analysis</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/576/1*5oHLiOSwG5v4uQVk0dD5vA.png" /></figure><p>We see that Clouds , Clear &amp; Dust weather condition is the most dominating weather condition, we can classify some of the less common which are similar such as mist, fog ,smoke and Dust as Atmosphere</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/e8451d0b329339030aca3fd49f98c445/href">https://medium.com/media/e8451d0b329339030aca3fd49f98c445/href</a></iframe><figure><img alt="" src="https://cdn-images-1.medium.com/max/576/1*xefiin9SZ4-ja7xCeApRlw.png" /></figure><p>Let’s see the impact of weather conditions on Transport trip completion during Night or Day period across all the years.</p><ul><li><strong>2018</strong></li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*8mZMNzj9snP8NJTA39fEzg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*L33ZCDPHFEVC-hxIOieBCQ.png" /><figcaption>Transport Type</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/799/1*AnVlK2-3jzIuvw5aZsGGjQ.png" /></figure><ul><li><strong>2019</strong></li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*JIvQgQwQC3jcHf2fLBgMTw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4hngYF3fAD3sBoZi700x9A.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/783/1*48GuxCuqqAQy-csAuRR-7w.png" /></figure><ul><li><strong>2020</strong></li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*BEMlToJkSUagKpxzm_WQwA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*GyTyMefNF__0xHSQ0fyYRw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/775/1*4oU0hAQMfdYhQgOolSXCbQ.png" /></figure><h4>Insights Generated</h4><p>After analyzing the weather condition across the years to determine the impact on trips completion, the following information was observed;</p><ol><li>Given that 2018 had the highest number of trips completed, the trip completion rate was evenly distributed for both day&amp; night period all the weather condition with Clear &amp; Clouds weather conditions having the highest proportion of trips completed which can be term as <strong>good weather</strong>. A focus on what can be term as bad weather which can cause possible disruption in trips completionRain &amp; Thunderstorm, we see that the number of trips completed is less compared to the good weather condition, however, the trips completion rate was also evenly distributed for both periods of day and night.</li><li>In 2019, a focus on the bad weather condition Rain &amp; thunderstorm when compared to the trips completion rate for 2018, we observed they vary differently in the trip completion rate where;</li></ol><ul><li>Fewer trips were completed at night during Raining weather condition and an uneven distribution of trips completion between day and night</li><li>More trips were completed at night during Thunderstorm weather condition and also unevenly distributed</li><li>Few trips were completed at Night during Atmosphere weather condition such as Fog , Mist , Smoke etc</li></ul><p>A Hypothesis we can draw is that there is a possibility that when there is Rain which is accompanied by Thunderstorm trip completion rate is drastically reduced during the day with a minute tradeoff of little increase in trip completion at night.</p><p>To either reject or accept the hypothesis we performed the following :</p><ul><li>Explore the raining season with domain knowledge to determine the impact on trips completion in 2018, 2019 &amp; 2020 if that can explain the decline in the trip completion experience</li></ul><p>According to the research <a href="https://www.climatestotravel.com/climate/united-arab-emirates/dubai">here</a>, it is important to note that Rainfall in Dubai is infrequent and does not last for a long period. It mostly rains during the winter period between December and March in the form of short downpours and an <strong>occasional thunderstorm.</strong></p><p>First, let’s explore the year that had Occassional Thunderstorm during the raining season</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/7d18d539e01f87c8514fb29a63fde6a0/href">https://medium.com/media/7d18d539e01f87c8514fb29a63fde6a0/href</a></iframe><figure><img alt="" src="https://cdn-images-1.medium.com/max/576/1*tD5xF6H95Vo4K8gqbsmJQQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*eunh8Xa1mKEzlqByCSq4gA.png" /><figcaption>The year 2019</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*K_8TCJoS_M3NpYYpHErZCQ.png" /></figure><p>It is interesting to know that there was no Occasional thunderstorm in the year 2018 at all against the corresponding year 2019 &amp; 2020 which experienced an occasional thunderstorm with 2020 having the highest impact on the condition. An interesting discovery which implies that we can accept the Null Hypothesis that Thunderstorm really have an impact on the transport trip completion.</p><p>We can further explore the windspeed as well as the pressure in the explanatory features to be definite on the type of thunderstorm that impacts trip completion more given that there are different types of Thunderstorm such as ;</p><ul><li>Thunderstorm with light rain</li><li>Thunderstorm with heavy rain</li><li>Thunderstorm with light drizzle</li><li>Thunderstorm with heavy drizzle etc. as explained <a href="https://openweathermap.org/weather-conditions">here</a></li></ul><p>This project wouldn’t be drilling down and exploring further. However, with the insights gathered we can Engineer a binary feature which checks if there was a thunderstorm or not for our model to learn</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/a4b2bba0264b343fe429c19b9c181b02/href">https://medium.com/media/a4b2bba0264b343fe429c19b9c181b02/href</a></iframe><p><strong>Data Modelling</strong></p><p>This is the phase where a Machine learning model is trained using all the explanatory features to make a prediction for the target label Trips. Trips is selected as the target label given that the goal of the project is to predict disruption of Dubai public transport which can be interpreted as the impact of weather on commutation which can be inferred to be the Public transport trips completion.</p><p>The evaluation metrics to be used to evaluate the performance of the model is R_squared to help determine how the explanatory features are significant in explaining the variability of the target. Before the model is fitted on the data necessary feature transformation was performed on the data which include the following though not exhaustive;</p><ol><li>Filling Missing Informations</li><li>Encoding Categorical variables</li><li>Dropping features etc.</li></ol><p>The data modeling involved the following phases ;</p><ul><li>Importing important libraries required for modeling</li></ul><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/d4f8d8ea5cdc969e7ba3c7b3fea02ffe/href">https://medium.com/media/d4f8d8ea5cdc969e7ba3c7b3fea02ffe/href</a></iframe><ul><li>Splitting our dataset in Training and validation set to be used to select our baseline model</li></ul><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/07289afe073abf7844130715a9165157/href">https://medium.com/media/07289afe073abf7844130715a9165157/href</a></iframe><ul><li>Cross-validation</li></ul><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/566a667f59cc650ee1e72a5c769a5375/href">https://medium.com/media/566a667f59cc650ee1e72a5c769a5375/href</a></iframe><pre>LR: 0.7724538353110713 0.004688685956395516<br>KNN: 0.06913799588491851 0.013173344887201933<br>CART: 0.9999962241169195 3.873468729874052e-06<br>GB: 0.9860423570101974 0.001426326504846694</pre><ul><li>Select Baseline model</li></ul><p>From the result of the cross-validation, we can see that CART &amp; GB had better accuracy to enable us to select a baseline model to improve on. We selected Gradient Boosting as our baseline model for this project.</p><ul><li>Perform Prediction &amp; Evaluate on Baseline Model</li></ul><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/9dcac2390072fc0f573252f49326c36b/href">https://medium.com/media/9dcac2390072fc0f573252f49326c36b/href</a></iframe><pre>0.9960754457850877</pre><blockquote>Machine Learning is an iterative process: Don’t expect it to work first time, so the above is not an exhaustive approach to all Data Science problems, so it is important to try out different ideas and evaluate the model performance. Andre Ng states that his approach to building machine learning software is threefold ; Start off with an idea, Implement the idea in code &amp; Carry out an experiment to conclude how well the idea worked as shown in the diagram below</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/729/1*UNJH2Bk3xu4PiwycZxrtiw.png" /><figcaption>Source: Andrew Ng Machine Learning Yearning book</figcaption></figure><p><strong>Deployment</strong></p><p>The Deployment of Machine Learning or Data Science Solution can vary from a web application, mobile application, storytelling in the form of Data visualization for stakeholders, or technical reports for a Manager or superior, so the deployment will be dependent on how it will be utilized.</p><p>The deployment required for the Hackathon &amp; this project is in the form of storytelling using a blog or article. A web application may be considered in the future as well as further analysis to discover insights that will be useful in training the machine learning model to improve the current performance of the model in making predictions for unseen future data.</p><p>Thank you for the time and don’t forget to clap.</p><p>To see more about this analysis, see the link to my Github available <a href="https://github.com/Opiano1/Udacity-DSND/tree/master/Capstone">here</a></p><p>You can connect with me on ;</p><p><a href="https://www.linkedin.com/in/opeyemifabiyi">LinkedIn</a></p><p><a href="https://twitter.com/Opiano_1">Twitter</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=86373a19e6bc" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How Severe Are Car Accidents in the US?]]></title>
            <link>https://fabiyiopeyemi94.medium.com/how-severe-are-car-accidents-in-the-us-c86af93ad050?source=rss-343ff6d33d3------2</link>
            <guid isPermaLink="false">https://medium.com/p/c86af93ad050</guid>
            <dc:creator><![CDATA[FABIYI OPEYEMI]]></dc:creator>
            <pubDate>Sat, 25 Apr 2020 19:23:55 GMT</pubDate>
            <atom:updated>2020-04-27T17:00:14.400Z</atom:updated>
            <content:encoded><![CDATA[<p><strong>Introduction</strong></p><p>Truly, Death is inevitable, however, some deaths can’t be wished, one of which is death resulting from a car accident. When a death occurs as a result of a car accident, family members are often left shocked, grief tormented, and unsure of the uncertainty that might happen next. For surviving family members of fatal car accidents, there was never a chance to say goodbye to their loved one, resulting in emotional pain and anguish, too.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/962/1*93J9eAm2zLFW-QGIgbm3hg.jpeg" /><figcaption>source: <a href="https://www.iii.org/article/background-on-no-fault-auto-insurance">https://www.iii.org/article/background-on-no-fault-auto-insurance</a></figcaption></figure><p>Can this trauma be avoided? A resonating echoed answer, yes, I believe Data science can help give insights on the time of the day, the period of the year as well as what factors drive accidents occurrence in different regions. All of these questions motivated me to use my skills to gain insights to provide answers that might be helpful towards reducing the lives lost to car accidents.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/333/1*ibhjGLQ1qqZPv5R8cU1-Gw.png" /></figure><p>Therefore, I used data from Kaggle’s US 2019 updated Car Accidents, to answer the questions posed.</p><p>The survey data covers over 3M+ records with 49 different features which help drive insights to different aspects of the subject. For the sake of this project, the following questions will be answered;</p><ol><li>What is the Time distribution of Accident by Severity?</li><li>What is the distribution of Accidents by Location?</li><li>What is Accident Severity base on Weather Conditions?</li></ol><h3>Part I: What is the Time distribution of Accident Occurrence?</h3><p><strong>Daily Accident Severity Analysis</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/864/1*TaU4rhWG2A1rseMBgD5_pA.png" /><figcaption>Distribution of Accident Severity per Day</figcaption></figure><p><em>We observed that there is a decline in the severity of accidents that occurred during weekends as against accident severity during the early weekdays of Monday &amp; Tuesday.</em></p><p>Similarly, exploring the hourly distribution of accidents occurrence let us see that most accidents occurred around 7–8 am and at the evening, between 16–17 hrs with an average of 1hr from the start of the accident to the end and a very minimal accident takes more than 5 hours to be cleared.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1019/1*QZ-M6Bom0iTccD8W62Mtog.png" /><figcaption>Distribution of Accident Severity per Hour</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Y-SzeMWpVQNaM1K1L_jCpQ.png" /><figcaption>Time length of Accidents before Clearance</figcaption></figure><p><strong>Monthly Accident Severity Analysis</strong></p><p>Out of curiosity, I explored the monthly distribution to determine the season accidents occurrences are at the maximum as shown below, it is interesting to see that there is an increase in the number of accidents towards the end of fall (Aug/Sept) to the beginning of Winter October had the highest.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1008/1*wOE0X72VUXOEPxoaX2VF6g.png" /><figcaption>Monthly Distribution of Accidents</figcaption></figure><p>We now explore Severity by month to determine the month that had the highest severity and during what time of the day do they occur.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/864/1*sHm19J4hVB_SlyLD4KqPoA.png" /><figcaption>Distribution of Accident Severity per Month</figcaption></figure><p>At severity 4 we also confirmed that most of the accidents occurred during the fall season with an obvious increase in <strong>October</strong> and <strong>December</strong>. An in-depth look into the 3 months to determine the time of the day those high severe accidents occur revealed that most of the accidents happened around 12 pm.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/432/1*JpVfEsvM1-YzruhTRnkXWA.png" /><figcaption>Focus mode on October-December</figcaption></figure><h3>Part II: What is the distribution of Accidents per Location?</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*btmT4tZXqSnSLPCLJ7I2Pg.png" /><figcaption>Sources of Accident Reports</figcaption></figure><p><em>As seen Majority of the sources of the data were from MapQuest</em></p><p><strong>Location Accident Analysis</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*d8WNPHYxZ3xH1tVB8SPbhg.png" /><figcaption>Distribution of Accidents Occurrence by States</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*smOVlJGFJTE4QozT8-VNsw.png" /><figcaption>Frequency Distribution of Accidents by State in the USA (December 2019)</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/576/1*HWZsSswov-gaz1tsmiSHgw.png" /><figcaption>Distribution of Accident Occurrences by City</figcaption></figure><p>From the above plots, we explore accidents occurrences in terms of accidents by states and cities, California and Houston respectively have the largest proportion of accident occurrences in the entire region of the United State, and the majority of the data source was from MapQuest with a 74% proportion.</p><h3>Part III: What is Accident Severity base on Weather Conditions?</h3><p><strong>Weather Condition Accident Severity Analysis</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-nsS87zpzRraVhELL_EEFA.png" /><figcaption>Distribution of Accident Severity per Weather condition</figcaption></figure><p>As seen across all severity, most accidents occurred during Clear and a little cloudy weather which are mostly the regular weather conditions. However, it is important to explore whether conditions that are considered to be bad to determine how the severity of the accidents increased such as; Fog, Light rain, Snow, Rain, and Heavy Rain. This, therefore, implies that we will check how the severity of the accident increased for this weather Condition.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ey8kLg4hK0zn7rtKBS4Nrw.png" /><figcaption>Distribution of Accidents Severity by poor weather condition</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/864/1*Lu39oio2SIr_m6PjsghTBw.png" /></figure><p>In the above analysis, severity status 3 &amp; 4 that are considered to be highly severe were combined to measure their proportional increase to bad weather, and as shown, the proportion of Severity of accidents (case 3 &amp; 4) increased linearly with weather conditions;</p><p>Fog 25%, Light Rain 36%, Rain 39%, Heavy Rain 40%, and Snow 41%</p><p><strong>Conclusion</strong></p><p>In this article, we took a look at US car accident severity analysis according to the Kaggle dataset updated as of December 2019.</p><ol><li>We see that most accidents occurred around 7–8 am and at the evening, between 16–17 hrs with an average of 1hr from the start of the accident to the end and very minimal accidents takes more than 5 hours to be cleared, also we also discovered that accidents occur mostly during the fall period in which 50% of highly severe accidents happen around 12 pm.</li><li>We see that in term of accidents by states and city, California and Houston respectively have the largest proportion of accident occurrence in the entire region of United State, and the majority of the data source were from MapQuest with a 74% proportion</li><li>Finally, we looked at the impact of variation in weather conditions on the severity of accidents occurrence and saw that there is a linear relationship between poor weather and accident severity.</li></ol><p>The findings here are observational, not the result of a formal study. So you can further explore the data to garner more insights;</p><blockquote>“If you torture the data long enough it will confess to anything “ Ronald H. Coase</blockquote><p>Thank you for reading.</p><p>To see more about this analysis, see the link to my Github available <a href="https://github.com/Opiano1/Data-Scientist-Nanodegree">here</a></p><p>You can connect with me on ;</p><p><a href="https://www.linkedin.com/in/opeyemifabiyi">LinkedIn</a></p><p><a href="https://twitter.com/Opiano_1">Twitter</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c86af93ad050" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[MY DATA SCIENCE LEARNING JOURNEY]]></title>
            <link>https://fabiyiopeyemi94.medium.com/my-data-science-learning-journey-dc20b4a4a7f5?source=rss-343ff6d33d3------2</link>
            <guid isPermaLink="false">https://medium.com/p/dc20b4a4a7f5</guid>
            <dc:creator><![CDATA[FABIYI OPEYEMI]]></dc:creator>
            <pubDate>Thu, 02 Jan 2020 08:18:03 GMT</pubDate>
            <atom:updated>2020-01-02T08:18:03.990Z</atom:updated>
            <content:encoded><![CDATA[<p>MY DATA SCIENCE LEARNING JOURNEY</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DYA-hvNToSsEmfA3mOExGg.png" /></figure><p>Yippy!!!!</p><p>You made it!</p><p>If you are reading this it’s implies you are either thinking of embarking the journey of honing the skills for the most sought after job which Harvard Business School Review said is the &quot;<a href="https://hbr.org/2012/10/data-scientist-the-sexiest-job-of-the-21st-century">Sexiest Job in the 21st Century</a>&quot;.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_rUyOw40CWrSMe-857UHuw.png" /></figure><p>Starting out in Data-science can really be overwhelming due to plethora of resources available and difficulty in navigating. Don’t worry you aren’t alone this was my state February, 2018 when I decided to switch career from an Engineering background to Data Science, and that’s why this article was written for all data enthusiasts who reached out to ask for guidance towards starting out in Data Science. After reading this I’m sure you will have all your fears and anxiety gone.</p><p>Is that a prophesy?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*JpeAbu2yhTBrzHFBLvA4Kg.png" /></figure><p><em>Just try read to the end!!!</em></p><p>I am not an expert because I’ve got so much to learn and neither a beginner having learnt a stack of tools to crunch and torture any data (Excel,PowerBI,SQL, Python, MachineLearning &amp;Deeplearning). So my experience in the last 20months , I will be sharing resources that has been pivotal to my learning curve in sequence of how I took them as a complete novice to getting my first Job as an Analyst.</p><p><strong>N:B</strong> <em>No single course or resource will make you proficient so it&#39;s a long journey and it&#39;s a Marathon not a SPRINT</em>.</p><p>As a complete beginner this was the route I followed;</p><p>Are you ready!!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6inVSeb0J0KwSHPxIhSvyw.png" /></figure><p><em>Let the drum roll begin!!!</em></p><p>1. Data analysis by PwC (5 courses) - I took the first 3 courses</p><p><a href="https://www.coursera.org/specializations/pwc-analytics">https://www.coursera.org/specianalyticslizations/pwc-a</a></p><p>-Learnt Excel ,PowerPoint &amp; Presentation</p><p>Excel still remain the underutilized tool for Data Analytics, so learning this tool will be helpful in your journey.</p><p>2. MIT python course</p><p><a href="https://ocw.mit.edu/courses/electrical-engineering-and-computer-science/6-0001-introduction-to-computer-science-and-programming-in-python-fall-2016/">https://ocw.mit.edu/courses/electrical-engineering-and-computer-science/6-0001-introduction-to-computer-science-and-programming-in-python-fall-2016/</a></p><p>I complemented this with 2 great books ; Python for Informatics &amp; Automating boring stuff with Python.</p><p>3. Descriptive &amp; Inferential Statistics course by Udacity</p><p><a href="https://www.udacity.com/course/intro-to-descriptive-statistics--ud827">https://www.udacity.com/course/intro-to-descriptive-statistics--ud827</a></p><p><a href="https://www.udacity.com/course/intro-to-inferential-statistics--ud201">https://www.udacity.com/course/intro-to-inferential-statistics--ud201</a></p><p>You can complement this with Practical Statistics for Data-scientist book.</p><p>4. Python for Data-science and Machine Learning on Udemy</p><p><a href="https://www.udemy.com/python-for-data-science-and-machine-learning-bootcamp/">https://www.udemy.com/python-for-data-science-and-machine-learning-bootcamp/</a></p><p>5. Microsoft Professional Certification for Data-science (14 courses)</p><p><a href="https://academy.microsoft.com/en-us/professional-program/">https://academy.microsoft.com/en-us/professional-program/</a></p><p>Microsoft will be retiring this program this year 31/12/2019, however you can still take the courses individually here ;</p><p><a href="https://www.edx.org/school/microsoft">https://www.edx.org/school/microsoft</a></p><p>Feel free to reach out to me on the sequence of how to take the courses. It&#39;s a long and challenging ride which you enjoy<br>6. Andrew NG Machine Learning Course on coursera</p><p><a href="https://www.coursera.org/learn/machine-learning">https://www.coursera.org/learn/machine-learning</a></p><p>7. TOOK A BREAK AND PARTICIPATED IN COMPETITIONS AND PROJECTS TO STRENGTHEN MY KNOWLEDGE OF LESSON LEARNT FROM SEVERAL COURSES</p><p>8. Feature Engineering</p><p><a href="https://www.udemy.com/course/feature-engineering-for-machine-learning/">https://www.udemy.com/course/feature-engineering-for-machine-learning/</a></p><p>9. Feature Selection</p><p><a href="https://www.udemy.com/course/feature-selection-for-machine-learning/">https://www.udemy.com/course/feature-selection-for-machine-learning/</a></p><p>10. Andrew NG Deep Learning course</p><p><a href="https://www.coursera.org/specializations/deep-learning">https://www.coursera.org/specializations/deep-learning</a></p><p>ROLE SPECIFIC COURSES TO STRENGTHEN MY SQL SKILLS</p><p>11. <a href="https://www.linkedin.com/learning/sql-essential-training-2018">https://www.linkedin.com/learning/sql-essential-training-2018</a></p><p>12.</p><p><a href="https://www.linkedin.com/learning/advanced-sql-for-data-scientists">https://www.linkedin.com/learning/advanced-sql-for-data-scientists</a></p><p>Most importantly PRACTISE! PRACTISE!! PRACTISE!!!</p><p>I understood most concepts taught in courses better after participating in competitions , so pick up datasets on ;</p><p>-Kaggle :</p><p><a href="https://www.kaggle.com/">https://www.kaggle.com</a></p><p>-Zindi :</p><p>zindi.africa/competitions<br>-Analytics Vidhya : <a href="https://datahack.analyticsvidhya.com/?utm_source=feed_navbar">https://datahack.analyticsvidhya.com/?utm_source=feed_navbar</a><br>Lastly one way to stay motivated and not loose enthusiasm is to join a data Community as you keep whetting those skills. My friend and I launched a Telegram data commuity (<a href="https://t.me/joinchat/KCKqnBa6HjOnCWM7F6-9jQ">https://t.me/joinchat/KCKqnBa6HjOnCWM7F6-9jQ</a>)<br> to help enthusiasts navigate their way, so feel free to join and also reach out when you need help.</p><p>I am always available to help in my capacity towards democratizing AI, Data-science and Machine Learning.</p><p>Kindly drop your claps.</p><p><a href="https://www.linkedin.com/in/opeyemifabiyi">https://www.linkedin.com/in/opeyemifabiyi</a></p><p>fabiyiopeyemi94@gmail.com <a href="https://twitter.com/I_am_OpiAno?s=09">https://twitter.com/I_am_OpiAno?s=09</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=dc20b4a4a7f5" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>