<?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 Giovanna Fracasso on Medium]]></title>
        <description><![CDATA[Stories by Giovanna Fracasso on Medium]]></description>
        <link>https://medium.com/@giovanna.fracasso?source=rss-fc56881b7ce5------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*PPs5J77MDNma0UBX</url>
            <title>Stories by Giovanna Fracasso on Medium</title>
            <link>https://medium.com/@giovanna.fracasso?source=rss-fc56881b7ce5------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 12:27:57 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@giovanna.fracasso/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[Exploring the potential of synthetic data generation in the area of software quality]]></title>
            <link>https://medium.com/@giovanna.fracasso/exploring-the-potential-of-synthetic-data-generation-in-the-area-of-software-quality-4fd3c5297516?source=rss-fc56881b7ce5------2</link>
            <guid isPermaLink="false">https://medium.com/p/4fd3c5297516</guid>
            <dc:creator><![CDATA[Giovanna Fracasso]]></dc:creator>
            <pubDate>Tue, 28 May 2024 15:21:51 GMT</pubDate>
            <atom:updated>2024-05-28T17:01:20.414Z</atom:updated>
            <content:encoded><![CDATA[<p>While working in a team focused on data, the need arose to deliver tasks in a shorter time, but what made it impossible was that we depended on another team to generate the data. It was then that I identified the possibility of using synthetic data generation as a possible solution.</p><p><strong>The Motivation</strong><br>The need to improve task delivery times without depending on other teams was the main motivation for exploring new tools. In addition, we wanted to simplify the validation of our results with stakeholders, making the process more efficient and secure.</p><p><strong>The hypotheses</strong><br>With this in mind, we formulated some hypotheses:<br>Can we generate masses of data independently?<br>Is it possible to use synthetic data in the analysis phase to facilitate validation with stakeholders?</p><p><strong>Discovering synthetic data</strong><br>While investigating solutions, we discovered synthetic data. This is artificially created data that mimics the patterns of real data. They are ideal for situations where using real data may be expensive or raise privacy concerns. Using synthetic data allows us to develop and test algorithms and models without compromising data privacy.</p><p><strong>Why use synthetic data?</strong><br>Synthetic data has proven to be a versatile tool, offering solutions in many situations where the use of real data is impractical, unfeasible or poses significant risks. They offer<br>- <em>Privacy and Security:</em> They avoid exposing sensitive data.<br>- <em>Variety and Scope:</em> They allow for the creation of varied and exceptional scenarios.<br>- <em>Cost and Time:</em> They reduce the time and cost of data collection and preparation.</p><p><strong>Getting to know SDV</strong><br>During our research, we learned about the Synthetic Data Vault (SDV), a Python library designed to generate tabular synthetic data while maintaining the statistical and structural characteristics of real data.</p><p><strong>Benefits of SDV</strong><br>Opting for SDV brings several advantages:<br>- <em>Privacy:</em> Guarantees the protection of sensitive data.<br>- <em>Flexibility and Applicability:</em> Adapts to different types of data and scenarios.<br>- <em>Easy to implement:</em> Simple to configure and use.<br>- <em>Good Documentation:</em> Excellent support and learning resources.<br>- <em>Generated Data Quality Validation:</em> Automatically checks the quality of synthetic data.</p><p><strong>How does SDV work?</strong><br>- <em>Real Data Input:</em> We input real data into the ML software.<br>- <em>Machine Learning:</em> The software automatically learns the patterns in the data.<br>- <em>Data Generation:</em> The software generates data that contains these patterns.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*E8hDtja9q5OxpeP_OfKdPg.png" /></figure><p>Despite the advantages, we identified that, in our context of sensitive financial data, the security of using the tool still raised doubts. However, I believe that for other teams that don’t deal with such sensitive data, SDV can be extremely useful.</p><p>Exploring the use of synthetic data can be a competitive differentiator for teams. I recommend that professionals in the field consider this approach to optimize their processes and ensure more comprehensive and secure testing.</p><p>I would especially like to thank <a href="https://medium.com/u/111556840812">Taluna Mendes</a> for her valuable help in the research and discussions that guided us on this journey.</p><p>If you would like to know more about SDV, below are useful links<br>- <a href="https://sdv.dev/">https://sdv.dev<br></a>- <a href="https://docs.sdv.dev/sdv">https://docs.sdv.dev/sdv<br></a>- <a href="https://datacebo.com/blog/fake-to-synthetic-ml/">https://datacebo.com/blog/fake-to-synthetic-ml/</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4fd3c5297516" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>