<?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 Jake Goodman on Medium]]></title>
        <description><![CDATA[Stories by Jake Goodman on Medium]]></description>
        <link>https://medium.com/@jakegoodman_49535?source=rss-a13c8a63ec95------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*F86jCbfnX_zhV69kZIj05w.jpeg</url>
            <title>Stories by Jake Goodman on Medium</title>
            <link>https://medium.com/@jakegoodman_49535?source=rss-a13c8a63ec95------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 04:02:04 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@jakegoodman_49535/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[Flask: A Data Analyst Friendly Web App Framework?]]></title>
            <link>https://medium.com/@jakegoodman_49535/flask-a-data-analyst-friendly-web-app-framework-c56a6d0a9b2a?source=rss-a13c8a63ec95------2</link>
            <guid isPermaLink="false">https://medium.com/p/c56a6d0a9b2a</guid>
            <category><![CDATA[data-analyst]]></category>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[python]]></category>
            <category><![CDATA[flask]]></category>
            <category><![CDATA[data-analytics]]></category>
            <dc:creator><![CDATA[Jake Goodman]]></dc:creator>
            <pubDate>Tue, 02 Apr 2024 01:53:56 GMT</pubDate>
            <atom:updated>2024-04-02T01:53:56.372Z</atom:updated>
            <content:encoded><![CDATA[<h3><strong>Background</strong></h3><p>Let’s go back a year or so where it was just another day where our data analyst team was fielding some data related requests in a Slack channel. On this day we happened to get a few similar requests that entailed some very simple adhoc database updates on a table where we manage some email addresses for notifications. Again, these updates are super simple and not that time consuming but the amount of these requests we were getting led me to the classic internal commentary, “there has to be a better way!”.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fgiphy.com%2Fembed%2FyAnH8autwgI3fKVNME%2Fvideo&amp;display_name=GIPHY&amp;url=https%3A%2F%2Fgiphy.com%2Fclips%2Fneonrated-neon-rated-how-to-blow-up-a-pipeline-yAnH8autwgI3fKVNME&amp;image=https%3A%2F%2Fmedia3.giphy.com%2Fmedia%2FyAnH8autwgI3fKVNME%2Fgiphy.gif&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=giphy" width="480" height="270" frameborder="0" scrolling="no"><a href="https://medium.com/media/6be793d07f066eff42674169ad6e2414/href">https://medium.com/media/6be793d07f066eff42674169ad6e2414/href</a></iframe><p>I work for a SaaS company so in an ideal world this could be a configuration/UI managed within our product but our engineers have<strong> way more</strong> important stuff to work on, so no need to even bother them with a request like this. Additionally, and not to spoil the ending 😉, there’s likely to be other use cases for this kind of framework beyond this specific use case.</p><p>When trying to think of potential solutions, I remembered when I had previously played around with <a href="https://flask.palletsprojects.com/en/3.0.x/">Flask</a> the previous year for a internal company hackathon and found it to be pretty intuitive since I’ve worked with Python for a bit and I had the most basic understanding of HTML/CSS (read as very little).</p><p>For those that aren’t familiar, Flask is a very basic, lightweight, web app framework for Python. There are countless tutorials out there (like this <a href="https://www.digitalocean.com/community/tutorials/how-to-make-a-web-application-using-flask-in-python-3">one</a>), so I won’t spend much time on how to use Flask but rather keep it high level and what our use cases for it were.</p><p>You’ll learn much more in any tutorial, but just to show how easy it is to get started, all you need is the below code (and Flask package) to get a local webpage running!</p><pre>from flask import Flask <br><br>app = Flask(__name__)<br><br>@app.route(&quot;/&quot;)<br>def intro():<br>    return &quot;Hello, World!&quot;</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*joIBBbt6gzwrGh4tzeSUIA.png" /><figcaption>Side by side of flask app code and web page</figcaption></figure><p>Throughout the rest of this post, I’ll walk through our journey with getting an internal web app framework set up, some lessons learned, and how the use case changed (expanded) over time.</p><h3>Devops Prerequisite</h3><p>I’ll be making references to this later on but it’s worth mentioning first here that this whole journey would not be possible without a great, flexible Devops team (or something of the like). You can do all the local development in the world, however it doesn’t mean much unless you have a way of making it accessible to other users (<strong>in a secure way</strong>).</p><p>I’ll make plenty more references to how much our Devops team helped here as it extends beyond the above but thought it was important enough to have its own section early on in this post.</p><h3>First App</h3><p>As mentioned above, the first use case I had in mind was to create a UI that our implementation team could use on their own to manage these email addresses for certain notifications so our team didn’t have to act as the middlemen to make manual database updates. It’s important to note that this table lives outside of our application’s database as I could not recommend proceeding down a path where a data analyst team owns a UI that can make updates to production application databases (worst case scenario scared me a little too much 😅).</p><p>This was probably the best first UI to try and build since all that was needed was a dropdown to select the customer (which selected the corresponding database) and then a table where the user could add, edit, or remove records. I don’t pretend this is anything more than just exposing this database table to more users, but it was the perfect place to start in creating more self-service tools!</p><p>Obviously I’m glossing over all the technical implementation details since there’s so much Flask content out there, however I will at least jot down some lessons learned to hopefully make the implementation a little easier and more user-friendly since us data analysts rarely have in-depth FE engineering experience 😜.</p><ul><li>Unless you have a lot of CSS experience, just go ahead and use <a href="https://pythonhosted.org/Flask-Bootstrap/">Flask Bootstrap</a> (<a href="https://www.techwithtim.net/tutorials/flask/flask-adding-bootstrap">tutorial</a>), which is a library that helps with automatically including web page designs so you don’t have to create custom ones. There will certainly be conflicting opinions from FE engineers, but it’s what allowed me to move pretty quickly from the start.</li><li>Further on the above bullet, I’d definitely recommend using template inheritance for your styles. This is allows all the main styling to be done on a base.html page and lets child pages inherit that styling. It’s been such a blessing having that set up as we’ve had more teammates add onto this framework.</li><li>If you have any tables that users interact with, I found this <a href="https://datatables.net/">DataTables</a> library to be pretty easy to use that came with pretty good default styling and pre-built features (like searching and exports).</li></ul><p>Once I was able to finish all this development locally, I was finally ready to have it deployed and made accessible to users (enter Devops…)</p><p>Like I previously mentioned, we’re very lucky to have a great Devops team, so all they really needed from us was a repository with this code and a Dockerfile and they’d take care of the rest. Easy enough right? Well, if you’re like me, you’d maybe heard of a Dockerfile before but you’ve never done anything with it. Just like Flask, and probably even more so, there’s so much documentation out there on Docker, but the gist is a Dockerfile is just instructions on how to run an application. So in our case, I just needed the Dockerfile to pull down a certain python version, tell it where my app is, install the necessary python packages, and run the app. Below is what the core of our Dockerfile looks like for example. For more technical implementation details, this <a href="https://www.freecodecamp.org/news/how-to-dockerize-a-flask-app/">tutorial</a> is pretty decent.</p><pre>FROM python:3.11-slim-bullseye<br>WORKDIR /app<br>COPY ./requirements.txt /app<br>RUN pip install -r requirements.txt<br>COPY . .<br>EXPOSE 5000<br>ENV FLASK_APP=app.py<br>CMD [&quot;flask&quot;, &quot;run&quot;, &quot;--host&quot;, &quot;0.0.0.0&quot;]</pre><p>Once our Devops team had the above, and some other details, they were able to get the application deployed and running on a URL (accessible based on certain VPN permissions) 🚀.</p><p>Even while knowing this app was about as basic as it gets, I thought it was one of the coolest things seeing it deployed for everybody to be able to use. From a UI perspective, I’m very much used to building out different Tableau views but to be able to build a UI that enabled other users to actually make updates to a dataset they’ve never had access to shed light that there was much more we could accomplish with this kind of framework…</p><h3>Second App</h3><p>“Second App”… you’re probably thinking, oh my gosh, is he really just going to give a narrative for every app they’ve built?? Definitely not, but this section will highlight a flaw in how this approach would scale.</p><p>After building the first app, I was on a bit of a high in thinking of all the different use cases I could use this kind of framework for. I immediately started building the next app which was a bit similar to the first app in that it’d be a simple-ish UI around exposing a new dataset to users that’d let them manage some Customer → Salesforce alignments. The alternative to this kind of UI would involve some kind of process around a messy excel document and our team making manual database updates. Being that it was so similar to the first app, it didn’t take me long to get it locally developed and off to our Devops team so they could help deploy it.</p><p>Our Devops team is great, so they helped get this deployed and like the first app, we quickly saw how valuable something like this could be.</p><p>After this app, it was very clear we’d have other ideas from myself and the rest of the team. In sharing that with our Devops team, they shed light on the flaw of this approach… it ain’t scalable. Not only did this deployment take quite a bit of time for Devops, but it was also costing our company more money for each additional app since they were on their own instances. That’s when our Devops team asked if we could instead have each of these services just be separate routes of the same app as opposed to whole different apps themselves. This would mean we’d only have one app to host/maintain which would be a win for decreased cost and decreased development time. Being pretty novice in the web development world, I wasn’t exactly sure how we’d refactor for that but after just a little research, I come to find out that’s exactly what <a href="https://flask.palletsprojects.com/en/3.0.x/blueprints/">Flask Blueprints</a> is for!</p><h3>Flask Blueprint</h3><p>Flask Blueprints is essentially a way modularize your application into different services/blueprints (pretty solid <a href="https://www.youtube.com/watch?v=_LMiUOYDxzE">youtube tutorial</a>). For our examples, each of the first two apps would be separate services/blueprints under the same app which would subsequently allow them to be accessed like the below:</p><pre># Example 1 URL<br>https://internal-da-app.com/example_1<br><br># Example 2 URL<br>https://internal-da-app.com/example_2</pre><p>While subtle and maybe something that I should’ve thought about from the start, but this refactor was a game changer for us. Since we already had the infrastructure set up from one of the previous apps, we were able to refactor and add everything to that repository while our Devops team was able to just change the URL name to be something more generalized.</p><p>Being able to leverage the existing infrastructure meant that this could be a framework and process entirely managed by our data analyst team going forwards (thanks to the immense leg work already done by our Devops team). With that, we could develop as much and as fast as our own processes would allow while also giving our users a single site to access for all of our team’s internal tools.</p><h3>How has this been taken further?</h3><p>One of the really cool things about creating tools on this kind of framework is the ceiling of innovation is really just what’s possible on the web. That can obviously be a pretty low ceiling when you’re a data analyst dabbling in web development 🤣, but with ChatGPT out there, it can beat the alternative of the “I’m sorry, we’re not able to do that in Tableau” response.</p><p>With that, not biased at all, I think we’ve done some pretty cool things while admittedly using <strong>a lot of ChatGPT </strong>and other documentation to figure out how to extend the capabilities we’re able to provide here. Such as:</p><ul><li>Go Live Monitoring Tool: incorporated <a href="https://www.chartjs.org/">Chart.JS</a> to create visuals that help highlight real-time adoption after a customer goes live that was not previously possible with Tableau (real-time part).</li><li>SQL Generating Tool: like I mentioned in the beginning, I can’t recommend this kind of framework be used to make updates to a production database (as opposed to just selecting data), but there has been value in providing a UI that generates the necessary SQL script based on certain actions. Even though it’s not end-to-end automation, it still sped up the process to generate the necessary scripts at least.</li><li>Log Transparency Tool: there are some logs that are extremely painful to access and digest. Currently, only our team and engineering teams have access to these logs, but one of our teammates was able to create a UI that surfaces the relevant logs based on user input as it’s able to make specific API calls from that input to retrieve and process the relevant logs so they’re more easily digestible.</li></ul><h3>Conclusion and what’s next</h3><p>We’ve seen use of this framework grow so much since it’s initial inception, so I’m sure we’ll continue to innovate and find new use cases. As a data analyst, it’s really cool to see us be able to make this kind of impact and progress our organization towards a more self-service model and less manual updates. I couldn’t recommend trying something out like this more as the impact has been huge, but also, it’s been a lot of fun getting more experience in this area.</p><p>Hope you enjoyed and happy building 👷!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/591/0*d6wplB9GfXZd6Lv2.jpg" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c56a6d0a9b2a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[So you’re a data analyst — what do you do?]]></title>
            <link>https://medium.com/@jakegoodman_49535/so-youre-a-data-analyst-what-do-you-do-951ee3200b94?source=rss-a13c8a63ec95------2</link>
            <guid isPermaLink="false">https://medium.com/p/951ee3200b94</guid>
            <category><![CDATA[data]]></category>
            <category><![CDATA[analytics]]></category>
            <category><![CDATA[careers]]></category>
            <category><![CDATA[data-analyst]]></category>
            <dc:creator><![CDATA[Jake Goodman]]></dc:creator>
            <pubDate>Mon, 11 Dec 2023 00:32:08 GMT</pubDate>
            <atom:updated>2023-12-11T00:32:08.273Z</atom:updated>
            <content:encoded><![CDATA[<h3>So you’re a data analyst — what do you do?</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*F7xMZ844vIDJWW7WBybrPw.png" /></figure><p>“What do you at work?” is a question I feel like I shouldn’t struggle at answering after being in my current company for 2+ years and in the data world for 5+ years. This question usually comes from friends who aren’t in data or tech, but there is even still some struggle if I were asked by someone who is more familiar with what a data analyst might do.</p><p>Initially I felt bad about not being able to concisely answer this question but after thinking about it a bit more, and hopefully not just over rationalizing, I think I can spin it as a positive. The answer to that question really depends on the week you asked me. Especially being at a relatively smaller company currently (~350 people), our data analyst team plays a very flexible role. For example, this past week, I spent a lot of time developing a script that processes data and includes it an email as part of a new DS feature our org is working on. The previous week was dedicated to troubleshooting and resolving some issues in our DE pipeline. Next week I’ll be focusing on adding new metrics to our analytics pipeline. You get the gist… how do you categorize all these kinds of things into a sentence or two description? Going back to the “positive spin” reference, when thinking more about it, it makes me happy that I have such variety in my work that makes answering that question so tough. At the end of the day, if I really had to boil the job down to a sentence, I think it’d be that we use data to solve problems. An obviously generic description, but one that I think gets at the heart of the role which is we love to find solutions for any of kind of problem and usually we do so in a pretty quick and scrappy manner.</p><p>I’ve worked for two companies in my professional tenure so I’m not going to pretend every data analyst struggles with this question like I do, but for the reasons listed above, I kind of hope they do. I think the variety of problems I work on provides so much exposure to different areas of the company and often gives me the latitude to try out creative solutions (as in maybe something other than creating another dashboard 😉).</p><p>Circling back to the initial, “what do you do for work?” question… it’s obviously a valid question, but I’d actually rather answer “what did you do at work this week?”. As is with many jobs I’m sure, I think there are rarely the same two data analysts jobs out there across companies. Sure, we’re going to share similar technical skills but I’d think/hope the problems we’re solving vary quite a bit. I always love reading or hearing about what other data analysts are working on so keep them coming and I’ll try to contribute more by periodically answering that question of “what did I work on this week?”.</p><p>Happy problem solving 🤓</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=951ee3200b94" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>