<?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 Yasemin Nukan on Medium]]></title>
        <description><![CDATA[Stories by Yasemin Nukan on Medium]]></description>
        <link>https://medium.com/@yaseminnukan?source=rss-0a85a54a453b------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*a-iqqTH9zn_nrUB1NYQBKQ.jpeg</url>
            <title>Stories by Yasemin Nukan on Medium</title>
            <link>https://medium.com/@yaseminnukan?source=rss-0a85a54a453b------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 09:56:30 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@yaseminnukan/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[Machine Learning Model Evaluation: A Comprehensive Guide]]></title>
            <link>https://medium.com/@yaseminnukan/machine-learning-model-evaluation-a-comprehensive-guide-80b7442da312?source=rss-0a85a54a453b------2</link>
            <guid isPermaLink="false">https://medium.com/p/80b7442da312</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[model-evaluation]]></category>
            <category><![CDATA[data-science]]></category>
            <dc:creator><![CDATA[Yasemin Nukan]]></dc:creator>
            <pubDate>Fri, 12 Jan 2024 15:53:37 GMT</pubDate>
            <atom:updated>2024-01-12T15:53:37.606Z</atom:updated>
            <content:encoded><![CDATA[<p>Machine Learning (ML) has revolutionized the way we solve complex problems, make predictions, and automate decision-making processes. As the field continues to advance, the importance of robust model evaluation becomes increasingly evident. Properly assessing the performance of machine learning models is crucial to ensuring their effectiveness and reliability. In this blog post, we will delve into the key aspects of model evaluation, exploring various metrics, techniques, and best practices.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*r8ft15J81fn4U9AK" /><figcaption>Photo by <a href="https://unsplash.com/@markuswinkler?utm_source=medium&amp;utm_medium=referral">Markus Winkler</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Why is Model Evaluation Important?</p><p>Before we dive into the details of model evaluation, let’s understand why it’s so crucial in the realm of machine learning. The primary goal of any ML model is to make accurate predictions on new, unseen data. However, not all models are created equal, and their performance can vary based on various factors. Model evaluation helps us:</p><ol><li><strong>Select the Best Model:</strong> With multiple algorithms and hyperparameters to choose from, evaluating models helps identify the one that performs optimally for a given task.</li><li><strong>Optimize Hyperparameters: </strong>Fine-tuning a model’s hyperparameters is a critical step in improving performance. Model evaluation aids in finding the right combination of hyperparameters for enhanced accuracy.</li><li><strong>Assess Generalization: </strong>A good model should generalize well to new, unseen data. Model evaluation provides insights into how well the model is likely to perform in real-world scenarios.</li></ol><h4>Key Metrics for Model Evaluation:</h4><ol><li><strong>Accuracy:</strong> The most straightforward metric, accuracy measures the percentage of correctly classified instances. While it’s a good starting point, accuracy alone may not be sufficient, especially in imbalanced datasets.</li><li><strong>Precision and Recall:</strong> Precision focuses on the accuracy of positive predictions, while recall measures the ability of the model to capture all positive instances. The trade-off between precision and recall is often managed using the F1 score, which combines both metrics.</li><li><strong>Area Under the Receiver Operating Characteristic (ROC-AUC):</strong> Particularly useful for binary classification problems, the ROC-AUC metric evaluates the trade-off between true positive rate and false positive rate.</li><li><strong>Mean Squared Error (MSE) and Mean Absolute Error (MAE):</strong> Commonly used for regression problems, MSE and MAE quantify the average squared and absolute differences between predicted and actual values, respectively.</li></ol><h4>Model Evaluation Techniques:</h4><ol><li><strong>Cross-Validation:</strong> To ensure that the model’s performance is not dependent on a particular subset of the data, cross-validation involves splitting the dataset into multiple folds, training the model on different combinations, and averaging the results.</li><li><strong>Confusion Matrix Analysis:</strong> A confusion matrix provides a detailed breakdown of true positive, true negative, false positive, and false negative predictions, offering insights into the model’s strengths and weaknesses.</li><li><strong>Learning Curves:</strong> Examining learning curves helps understand how a model’s performance evolves as the size of the training set increases, aiding in identifying overfitting or underfitting.</li></ol><h4>Best Practices for Model Evaluation:</h4><ol><li><strong>Use Multiple Metrics:</strong> Relying on a single metric can be misleading. Combine accuracy with precision, recall, and other relevant metrics to gain a comprehensive understanding of model performance.</li><li><strong>Consider Business Context:</strong> The choice of evaluation metric should align with the specific goals and requirements of the business problem at hand. What matters most — minimizing false positives, false negatives, or achieving a balanced approach?</li><li><strong>Address Class Imbalance:</strong> In scenarios where one class significantly outnumbers the other, accuracy alone can be misleading. Look at precision, recall, and F1 score to get a more accurate picture.</li><li><strong>Regularly Reevaluate Models: </strong>Machine learning models are not static entities. Periodically reassessing their performance on new data ensures they remain relevant and effective over time.</li></ol><p>In the dynamic and evolving field of machine learning, model evaluation stands as a critical pillar. By adopting a holistic approach that considers multiple metrics, employs robust evaluation techniques, and aligns with the business context, data scientists and machine learning practitioners can ensure that their models not only perform well in controlled environments but also demonstrate reliability and generalizability in real-world applications. As we continue to push the boundaries of what is possible with machine learning, a thoughtful and rigorous approach to model evaluation remains an indispensable part of the journey.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=80b7442da312" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Machine Learning vs Deep Learning]]></title>
            <link>https://medium.com/@yaseminnukan/machine-learning-vs-deep-learning-24ebc46beeff?source=rss-0a85a54a453b------2</link>
            <guid isPermaLink="false">https://medium.com/p/24ebc46beeff</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[deep-learning]]></category>
            <dc:creator><![CDATA[Yasemin Nukan]]></dc:creator>
            <pubDate>Fri, 05 Jan 2024 14:51:49 GMT</pubDate>
            <atom:updated>2024-01-05T14:51:49.063Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*gnrH6SAkVxznuTfr" /><figcaption>Photo by <a href="https://unsplash.com/@markuswinkler?utm_source=medium&amp;utm_medium=referral">Markus Winkler</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=24ebc46beeff" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Role of APIs in Data Science]]></title>
            <link>https://medium.com/@yaseminnukan/the-role-of-apis-in-data-science-054107301b89?source=rss-0a85a54a453b------2</link>
            <guid isPermaLink="false">https://medium.com/p/054107301b89</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[api]]></category>
            <dc:creator><![CDATA[Yasemin Nukan]]></dc:creator>
            <pubDate>Thu, 21 Dec 2023 22:07:29 GMT</pubDate>
            <atom:updated>2023-12-21T22:07:43.741Z</atom:updated>
            <content:encoded><![CDATA[<p>The Role of APIs in Data Science APIs (Application Programming Interfaces) play an important role in data science by facilitating access to and integration of various data sources and services into data analysis workflows.<br>In this blog, we discuss the importance of APIs in data science and how they enable data scientists to extract, process, and analyze data more efficiently.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*PycxjzzTXbui1l_7" /><figcaption>Photo by <a href="https://unsplash.com/@douglasamarelo?utm_source=medium&amp;utm_medium=referral">Douglas Lopes</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><h4>What is an API?</h4><p>An API acts as a bridge between different software applications or services, allowing them to communicate and exchange information.<br> It defines a set of rules and protocols that govern how different software components can interact.</p><h4>Using APIs for Data Collection</h4><p>Data collection is an important task in data science, and APIs provide a structured way to access data from a variety of sources, such as websites, social media platforms, and databases. Simplifies this process by providing easy access.For example, popular APIs such as the Twitter API, Instagram API, and Google Maps API allow data scientists to extract real-time data about user activity, posts, and geographic information.</p><h4>Integrating APIs into Data Analysis Workflows</h4><p>Once data is collected, APIs help you seamlessly integrate it into data analysis workflows. APIs allow data scientists to connect to specific services, libraries, or tools that can perform advanced data analysis tasks.<br> For example, APIs provided by machine learning frameworks such as TensorFlow and PyTorch allow data scientists to leverage pre-trained models and efficiently perform complex calculations on large datasets.</p><h4>Improving Data Processing with APIs</h4><p>APIs also play an important role in data preprocessing and data processing tasks. Many APIs provide data transformation, cleansing, and enrichment functionality. For example, APIs provided by popular data analysis libraries such as Pandas and NumPy enable data manipulation, transformation, and cleansing operations in a user-friendly manner.</p><h4>Collaboration and Reusability</h4><p>APIs facilitate collaboration among data scientists and enable teams to reuse code and models. By building APIs around data analytics code or machine learning models, data scientists can create reusable components that can be shared across projects. This reduces redundancy, improves code maintainability, and fosters collaboration among team members.</p><h4>Real-time Data Analysis</h4><p>APIs facilitate real-time data analysis by allowing data scientists to stream and process generated data. Real-time APIs allow users to push or retrieve data instantly, allowing for real-time analysis, monitoring, or alerting based on the incoming data stream. This is especially valuable in areas such as finance, IoT, and social media analytics.</p><h4>Conclusion</h4><p>APIs are a fundamental part of the data science ecosystem and offer countless benefits to data scientists. It simplifies data collection, enables integration with a variety of services and tools, improves data processing capabilities, fosters collaboration, and facilitates real-time data analysis. As data science evolves, APIs will play an increasingly important role in enabling data scientists to maximize the value and insights from their data.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=054107301b89" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to Decide on the Right Data Visualization]]></title>
            <link>https://medium.com/@yaseminnukan/how-to-decide-on-the-right-data-visualization-9b8a49c52d71?source=rss-0a85a54a453b------2</link>
            <guid isPermaLink="false">https://medium.com/p/9b8a49c52d71</guid>
            <category><![CDATA[data-visualization]]></category>
            <category><![CDATA[data-science]]></category>
            <dc:creator><![CDATA[Yasemin Nukan]]></dc:creator>
            <pubDate>Thu, 14 Dec 2023 13:50:56 GMT</pubDate>
            <atom:updated>2023-12-14T23:45:03.056Z</atom:updated>
            <content:encoded><![CDATA[<p>Data visualization plays a crucial role in communicating insights and patterns from complex data sets. However, with the wide array of visualization options available, choosing the right one for your specific data can be a daunting task. In this article, we’ll explore a step-by-step approach to help you decide on the most effective data visualization technique.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*BUyDAN5YNoVVZIbV" /><figcaption>Photo by <a href="https://unsplash.com/@goumbik?utm_source=medium&amp;utm_medium=referral">Lukas Blazek</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><ul><li><strong>Understand your data:</strong><br>Before deciding on a visualization, take the time to understand your data. Identify the type of data you have (categorical, numerical, hierarchical, etc.), the relationships between variables, and the key insights you want to convey. This understanding will lay the foundation for choosing the appropriate visualization.</li></ul><p>Let’s assume you have a dataset containing information about sales figures across different regions over a period of time. We’ll begin by loading the data and performing some initial analysis:</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/b2f2bca5bd2189c49e9d25080bcd5cbe/href">https://medium.com/media/b2f2bca5bd2189c49e9d25080bcd5cbe/href</a></iframe><p>After loading the data, the head() function helps to get a glimpse of the dataset by displaying the first few rows. describe() provides statistical summaries like mean, standard deviation, min, max, etc., for numerical columns. info() gives information about data types and missing values.</p><p>Once you understand the structure and content of your data, you can choose appropriate visualization techniques. For instance, if you want to visualize the trend of sales over time for different regions, a line plot might be suitable:</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/f4783c923a6846f34d79d6e0b62e7427/href">https://medium.com/media/f4783c923a6846f34d79d6e0b62e7427/href</a></iframe><p>This code creates a line plot using matplotlib, displaying the sales trend across different regions over time. Replace &#39;Date&#39; and &#39;Sales&#39; with the respective column names from your dataset.</p><p>Understanding the nature of your data — whether it’s categorical, numerical, time-series, etc. — will guide you in choosing the appropriate visualization technique to effectively communicate insights and patterns.</p><ul><li><strong>Define your objectives:</strong><br>Clearly define what you want to achieve with your data visualization. Are you trying to highlight trends, compare values, show the distribution, or represent proportions? Understanding your objectives will guide you towards the appropriate visual representation.</li></ul><ol><li><strong>Highlight Trends:</strong> If your goal is to showcase trends or patterns over time or across different categories, suitable visualizations include line charts for temporal trends, bar/column charts for categorical comparisons, or area charts for cumulative trends.</li><li><strong>Compare Values: </strong>When comparing values across different categories or groups, bar charts, grouped bar charts, stacked bar charts, or box plots are effective choices. They allow for clear comparisons between different entities.</li><li><strong>Show Distribution: </strong>Representing the distribution of data is often done using histograms for continuous variables or box plots/violin plots to display the spread and central tendency of the data.</li><li><strong>Represent Proportions: </strong>Visualizations like pie charts or stacked bar/column charts are useful for representing proportions or percentages within a whole. They show how parts relate to the whole.</li><li><strong>Identify Relationships or Correlations:</strong> Scatter plots are excellent for showcasing relationships or correlations between two variables. Heatmaps and correlation matrices can also help visualize relationships in larger datasets.</li><li><strong>Display Geographic Patterns: </strong>Maps, choropleth maps specifically, are great for displaying geographical data or patterns across different regions.</li><li><strong>Visualize Hierarchical Structures:</strong> Tree maps, sunburst charts, or dendrograms are useful for visualizing hierarchical data structures.</li></ol><p>Before selecting a visualization type, clearly defining your objective helps in choosing the most appropriate and effective visualization technique to convey the desired message from your data.</p><ul><li><strong>Consider the message:</strong><br>Think about the message you want to convey to your audience. Do you want to emphasize patterns, compare values, or show trends over time? Different visualization techniques excel at different messages. For example, line charts are effective for showing trends, bar charts for comparisons, and scatter plots for exploring relationships.</li></ul><ol><li><strong>Emphasizing Patterns: </strong>If your goal is to highlight patterns within your data, techniques like line charts, area charts, or heatmaps can be effective. They excel at showcasing trends or patterns over time or across categories. Line charts are great for continuous data trends, while heatmaps can reveal patterns in larger datasets.</li><li><strong>Comparing Values:</strong> When the main objective is to compare values between different categories or groups, bar charts, grouped bar charts, stacked bar charts, or box plots work well. They allow for easy comparison and identification of differences or similarities.</li><li><strong>Showing Trends Over Time:</strong> If the primary message is to showcase trends or changes over a period, line charts are the go-to choice. They visualize trends or fluctuations in data points over time, making them ideal for time-series data analysis.</li><li><strong>Exploring Relationships:</strong> When exploring relationships or correlations between two or more variables, scatter plots are invaluable. They help identify patterns, clusters, or trends between variables, revealing the strength and direction of relationships.</li><li><strong>Highlighting Distribution:</strong> To emphasize the distribution of data, histograms, box plots, or violin plots are effective. They illustrate the spread, central tendency, and skewness of the data distribution.</li><li><strong>Comparing Proportions or Parts to Whole:</strong> Pie charts, stacked bar/column charts, or treemaps are useful for representing proportions or percentages within a whole. They visualize how parts contribute to the entirety.</li></ol><p>By considering the specific message you want to convey, you can select the most appropriate visualization technique that effectively communicates that message to your audience. Each technique excels at highlighting different aspects of the data, so choosing the right one ensures your audience grasps the intended information clearly and accurately.</p><ul><li><strong>Select the right visualization type:</strong><br>Based on your understanding of the data, objectives, and desired message, choose the most appropriate visualization type. Some commonly used types include bar charts, line charts, scatter plots, pie charts, area charts, and heatmaps. Consider the strengths and limitations of each type and select the one that best represents your data and conveys your message effectively.</li></ul><p>Let’s consider a scenario where the objective is to compare the sales performance of different products over time for a retail business. The desire is to emphasize trends and variations in sales for each product category.</p><p>Given this scenario, a suitable visualization type could be a line chart. Here’s why:</p><p><strong>Strengths of Line Charts:</strong></p><ol><li><strong>Trend Emphasis: </strong>Line charts excel in showcasing trends and variations over time, making them ideal for visualizing changes in sales performance across different product categories.</li><li><strong>Clarity in Time-Series Data:</strong> They provide a clear representation of data points over a continuous period, allowing for easy identification of trends, peaks, or dips.</li><li><strong>Multiple Lines for Comparison: </strong>Multiple lines (one for each product category) on the same chart allow for direct comparisons between categories.</li></ol><p><strong>Limitations:</strong></p><ol><li><strong>Complexity with Many Categories:</strong> Line charts might become cluttered and hard to interpret if there are too many product categories to display.</li><li><strong>Not Suitable for Categorical Data:</strong> Line charts work best for showing trends in continuous data, so if the data is not time-based or lacks continuity, other visualizations might be more suitable.</li></ol><p>Here’s an example code snippet in Python using matplotlib to create a line chart for this scenario:</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/436152886f947efc958b97839a97e186/href">https://medium.com/media/436152886f947efc958b97839a97e186/href</a></iframe><p>This line chart will effectively display sales trends over time for various product categories, allowing for easy comparison and identification of sales patterns for each product.</p><ul><li><strong>Focus on clarity and simplicity:</strong><br>When selecting a data visualization, prioritize clarity and simplicity. Avoid cluttering the visualization with unnecessary elements that may confuse or distract the audience. Ensure that the chosen visualization effectively communicates the intended information, making it easy for the audience to understand and interpret.</li></ul><ol><li><strong>Limit Information:</strong> Avoid overcrowding your visualization with excessive data or elements. Focus on the essential information that supports your message or objective.</li><li><strong>Clear Labels and Titles:</strong> Use descriptive and concise labels for axes, data points, and legends. Ensure the title clearly reflects the content and purpose of the visualization.</li><li><strong>Appropriate Use of Colors: </strong>Choose a color scheme that enhances readability and comprehension. Avoid using too many colors or overly bright/pastel shades that might strain the audience’s eyes or make it hard to differentiate between elements.</li><li><strong>Sensible Scaling:</strong> Ensure that the scaling of axes in the visualization is logical and appropriate. Avoid misleading the audience by manipulating scales to emphasize certain aspects of the data unfairly.</li><li><strong>Simplify Text and Annotations: </strong>Minimize text clutter by using tooltips or selective annotations to highlight specific data points or trends. Use annotations sparingly to draw attention to key findings.</li><li><strong>Remove Unnecessary Chart Junk:</strong> Eliminate unnecessary gridlines, borders, or background elements that don’t contribute meaningfully to the visualization. Keep it clean and focused on the data.</li><li><strong>Test for Interpretability:</strong> Before presenting the visualization, ensure that someone unfamiliar with the data can interpret it easily. If it’s confusing or unclear, consider revising or simplifying the visualization.</li></ol><p>By prioritizing clarity and simplicity, you can create visualizations that effectively convey the intended information without overwhelming or confusing your audience. The goal is to make the data easily understandable and actionable.</p><ul><li><strong>Consider your audience:</strong><br>Keep your target audience in mind when deciding on a data visualization. Consider their familiarity with the subject matter and the level of complexity they can comprehend. If your audience is not data-savvy, choose visualizations that are intuitive and easy to understand. On the other hand, if your audience is more technically inclined, you can opt for more sophisticated visuals.</li></ul><ol><li><strong>Non-Data-Savvy Audience:</strong></li></ol><ul><li><strong>Choose Intuitive Visualizations: </strong>Opt for straightforward visualizations like bar charts, pie charts, or simple line charts that are easy to interpret without prior knowledge of data analysis.</li><li><strong>Avoid Complex Visuals:</strong> Minimize complex statistical graphs or intricate visualizations that might confuse or overwhelm non-technical audiences.</li><li><strong>Use Descriptive Titles and Labels: </strong>Ensure clear, descriptive titles, and labels to guide the audience through the information presented.</li></ul><p>2<strong>. Data-Savvy or Technical Audience:</strong></p><ul><li><strong>Consider Sophisticated Visualizations:</strong> For audiences comfortable with data analysis, you can use more complex visuals like scatter plots, histograms, or heatmaps that provide deeper insights into the data.</li><li><strong>Include Advanced Metrics or Trends: </strong>Incorporate more technical details, statistics, or advanced trends if your audience is familiar with these concepts.</li><li><strong>Allow Interactivity or Drill-Downs: </strong>Use interactive visualizations or tools that allow the audience to explore the data further based on their interests or expertise.</li></ul><p>3. <strong>Mixed Audience:</strong></p><ul><li><strong>Balance Complexity:</strong> If your audience includes both technical and non-technical individuals, aim for a balance. Use visualizations that convey the main message clearly without overwhelming those less familiar with data analysis.</li><li><strong>Provide Context and Explanations: </strong>Accompany your visualizations with contextual explanations or summaries to ensure everyone comprehends the key takeaways.</li></ul><p>Remember, the goal is to effectively communicate insights to your audience in a way that resonates with their knowledge and expertise. Adapting your visualization style to suit their familiarity with the subject matter enhances understanding and engagement.</p><ul><li><strong>Iterate and experiment:</strong><br>Don’t be afraid to iterate and experiment with different visualization types. Adjusting visualization parameters such as color, size, and layout can significantly impact the clarity and effectiveness of the visualization. Experiment with various options to find the one that best represents your data and meets your objectives.</li></ul><ol><li><strong>Start with Basic Visualization: </strong>Begin by creating a basic visualization that aligns with your initial understanding of the data and objectives. For example, if you’re considering a line chart, create a simple version using default settings.</li><li><strong>Experiment with Parameters:</strong> Try adjusting various parameters such as colors, line styles, marker types, axes scales, or chart sizes. Small changes in these elements can significantly impact the readability and clarity of the visualization.</li><li><strong>Explore Alternative Visualizations:</strong> Don’t stick to one type. Experiment with different visualization types that might suit your data better. For instance, if you started with a bar chart, try a line chart, scatter plot, or other types to see if they convey the information more effectively.</li><li><strong>Seek Feedback:</strong> Share your visualizations with colleagues or individuals who might represent your target audience. Gather feedback on what works well and what might be confusing or less clear. Their input can guide your adjustments.</li><li><strong>Refine Based on Insights: </strong>Incorporate the feedback and insights gained from experimentation. Refine the visualization by emphasizing the aspects that resonate with your audience while simplifying or adjusting elements that cause confusion.</li><li><strong>Iterate and Repeat: </strong>Continue the iterative process, making incremental changes based on feedback and your evolving understanding of the data and objectives.</li><li><strong>Document Your Findings: </strong>Keep track of the changes made and their impact on the effectiveness of the visualization. Documenting your experimentation process helps in understanding what works best for your specific dataset and objectives.</li></ol><p>Remember, there’s no one-size-fits-all approach to visualization. Iteration and experimentation allow you to fine-tune and customize your visualizations, ensuring they effectively communicate insights from your data.</p><p>Choosing the right data visualization is a critical step in effectively communicating insights from your data. By understanding your data, defining objectives, considering the message, and selecting the appropriate visualization type, you can ensure that your data visualization effectively conveys the intended information to your audience. Remember to prioritize clarity, simplicity, and the needs of your target audience. With practice and experimentation, you’ll become adept at selecting the right visualization technique for any given data set.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9b8a49c52d71" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Python Basics for Data Science]]></title>
            <link>https://medium.com/@yaseminnukan/python-basics-for-data-science-125370e13c0c?source=rss-0a85a54a453b------2</link>
            <guid isPermaLink="false">https://medium.com/p/125370e13c0c</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[python]]></category>
            <category><![CDATA[introduction-to-python]]></category>
            <dc:creator><![CDATA[Yasemin Nukan]]></dc:creator>
            <pubDate>Thu, 30 Nov 2023 15:03:01 GMT</pubDate>
            <atom:updated>2023-11-30T15:03:01.755Z</atom:updated>
            <content:encoded><![CDATA[<p>Hey there! Welcome to my blog on data science. In this post, we are going to dive into the basics of Python for data science. Python is an incredibly popular language in the field of data science due to its simplicity and extensive libraries. So, let’s get started!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*iGLUOPebAK_K589P" /><figcaption>Photo by <a href="https://unsplash.com/@cgower?utm_source=medium&amp;utm_medium=referral">Christopher Gower</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><h4>What is Python?</h4><p>Python is a high-level programming language known for its readability and simplicity. Its syntax emphasizes code readability, making it easier for beginners to grasp and understand. Python is not only beginner-friendly but also widely used in data science due to its extensive libraries and frameworks that facilitate various data manipulation, analysis, and visualization tasks.</p><h4>Python Basics:</h4><p>Let’s start by understanding some fundamental aspects of Python programming:</p><ul><li><strong>Variables and Data Types: </strong><br>In Python, you can declare variables without specifying the type explicitly. This is one of the benefits of a dynamically-typed language like Python. Some common data types include integers, floats, and strings. You can also create more complex data structures like lists, tuples, sets and dictionaries in Python.</li><li><strong>Control Flow:</strong><br>To control the flow of execution in Python, you can make use of loops and conditional statements. The ‘for’ loop allows you to iterate over a sequence of elements, while the ‘while’ loop allows you to repeat a block of code until a certain condition is met. Conditional statements, such as ‘if-else’, help you make decisions in your code based on specific conditions.</li></ul><p>First, I will give examples of data types. There are 5 different data types :</p><ul><li><strong>Integer:</strong> It holds number of non-limited length without a fraction.</li><li><strong>Float:</strong> It holds number with a fraction up to 15 decimal places.</li><li><strong>String:</strong> It holds sequence of characters. It holds a string of characters. Quotation marks are used at the beginning and end of the sentence.</li><li><strong>Boolean:</strong> It holds either True or False.</li><li><strong>None:</strong> It represents a null variable.</li></ul><p>You can check the data type of variable using the <strong>type()</strong> function.You can see examples below.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/b498be9b4321591f222b042469a4d626/href">https://medium.com/media/b498be9b4321591f222b042469a4d626/href</a></iframe><p>A list is an arranged grouping of elements that are either the same or different, separated by commas and enclosed in brackets [ ]. To access items from a list, we use the index number <strong>(0, 1, 2 …)</strong>. For instance,</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/551ef9a81fca19b5e2abd4dee5c4be4e/href">https://medium.com/media/551ef9a81fca19b5e2abd4dee5c4be4e/href</a></iframe><p>Similar to a list, a tuple is an ordered series of elements. Tuples vary solely in that they cannot be changed. Once formed, tuples cannot be changed.In Python, we use the parentheses ( ) to store items of a tuple. For example,</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/51a844785864a00fbb10882b8aa318b1/href">https://medium.com/media/51a844785864a00fbb10882b8aa318b1/href</a></iframe><p>A set is an arbitrary grouping of distinct objects. Values inside braces { } and separated by commas define a set. For instance,</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/d8e9a9a257d08922ef74c299086041b1/href">https://medium.com/media/d8e9a9a257d08922ef74c299086041b1/href</a></iframe><p>A Python dictionary is a sorted list of objects. Elements are kept in key/value pairs. In this case, keys are distinct identifiers connected to every value. Using this as an example,</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/34f39e5cdf91ec0f558673ac7363efad/href">https://medium.com/media/34f39e5cdf91ec0f558673ac7363efad/href</a></iframe><p>Let check lists, tuples, sets and dictionaries data types using the <strong>type()</strong> function.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/9a70e9bd461fdf0a9ff07525ca3b4762/href">https://medium.com/media/9a70e9bd461fdf0a9ff07525ca3b4762/href</a></iframe><p>I will explain the Python control flow part with examples to you in my next post. Don’t forget to stay tuned!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Jco_JK-RpTz07yxC" /><figcaption>Photo by <a href="https://unsplash.com/@mbaumi?utm_source=medium&amp;utm_medium=referral">Mika Baumeister</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><h4>Introduction to Data Science with Python:</h4><p>Now that we have covered the basics of Python, let’s explore how we can leverage Python for data science tasks.</p><p>Python plays a significant role in various stages of a data science workflow:</p><ul><li><strong>Data Manipulation:</strong><br>Python provides powerful libraries like NumPy and Pandas for handling data. NumPy allows efficient numerical computations, while Pandas provides data structures and functions for manipulating and analyzing data. These libraries are the backbone of any data science project.</li><li><strong>Data Visualization:</strong><br>Visualizing data is crucial in data science, and Python libraries like Matplotlib come to the rescue. Matplotlib helps you create visually appealing plots, charts, and graphs, making it easier to understand and communicate insights effectively.</li></ul><p>In this post, we covered the Python basics for data science. We explored Python as a popular language in the field of data science due to its simplicity and extensive libraries. We discussed variables, data types, control flow, and introduced some key libraries for data manipulation and visualization.</p><p>Python is just the tip of the iceberg when it comes to data science. There is so much more to learn and discover. So, why wait? Start exploring the amazing world of data science with Python today!</p><p>I hope this post provides you with a good introduction to Python basics for data science. Stay tuned for more exciting topics in the field of data science. Happy coding!</p><h4><strong>Who am I?</strong></h4><p>A passionate and recent Computer Engineering graduate with a specialization in Data Science, particularly in Machine Learning and Deep Learning applications. My academic journey culminated in a pivotal project focusing on “Distance Education Deep Sentiment Analysis During Pandemic with Data Augmentation,” a critical exploration into leveraging advanced data methodologies to understand and enhance remote learning experiences. I want to share with you what I researched. Knowledge grows when shared!</p><p>Follow me on Github:</p><p><a href="https://github.com/yasonukan">yasonukan - Overview</a></p><h4>References:</h4><ul><li><a href="https://www.learnpython.org/">Learn Python - Free Interactive Python Tutorial</a></li><li><a href="https://www.w3schools.com/python/">Python Tutorial</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=125370e13c0c" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>