From Jupyter Notebooks to Viral Medium Posts: My Journey as a Data Science Evangelist

In December 2020, I joined Datapane as a freelance data science evangelist, stepping into a small, remote UK startup with a powerful yet niche product. The challenge they were tackling was something many of us in the data science community were feeling acutely: how do you share your in-depth Jupyter notebook analysis with the world?

Jupyter notebooks are fantastic for deep-dive analysis, but they were, and in many ways still are, a closed ecosystem. You could do brilliant work, create intricate plots, and build interactive widgets, but if you wanted to share that with a wider audience, a blog post was a static, lifeless medium.

Datapane had a unique solution:

a platform that would allow you to host your notebooks, rendering them as interactive HTML pages that could be embedded anywhere. This wasn’t just a simple screenshot; it was a living, breathing article where readers could interact with the plots and widgets you had created.

My mission was to bridge the gap between this innovative technology and the data science community.

The role was unlike anything I had done before. My background was in data science, not marketing, but I was eager to experiment and see what I could accomplish. My job was to use the product, provide feedback, and, most importantly, create content that showcased its power. This meant doing deep analysis myself and then writing compelling Medium articles about the process.

The Two-Pronged Approach to Content Creation

My content strategy was a blend of direct product evangelism and a more subtle, value-driven approach. On one hand, I had to showcase what Datapane could do directly. I would take a complex analysis, use their product to publish it, and then write a Medium post detailing the process. This was about demonstrating, in real-time, how Datapane solved a problem that other tools couldn’t.

On the other hand, I created more general-interest articles that would attract a wider audience and then subtly redirect them to Datapane.

The whole idea was that I would create that article and then through referrals or, you know, redirections to Datapane's product, people would come in.

One of my most successful articles followed this second strategy: an in-depth guide to the top four Python libraries for creating amazing visualizations. This wasn't a simple listicle; it was a deeply researched, code-intensive piece of content that required significant effort.

I had to do a lot of research to figure out what works, what doesn't work, then bringing that research and then, you know, try those products out, try those libraries out, create some really interesting visuals, interesting dashboards out of it.

This wasn't an era of AI-generated code. Each line had to be meticulously written and tested. I experimented with complex libraries like Alteryx and Plotly to build interactive dashboards with filters and zoom functionality. I was building working examples, not just talking about them. The final articles were high-quality, actionable guides that provided genuine value to the reader.

This approach paid off in a big way. My articles weren’t just read; they were saved, shared, and referenced by thousands of people. I was getting thousands of views and thousands of readers, and many people were actually saving my article for reference to their work. This was a powerful validation that my content was making an impact and generating real value.

Beyond the Content: Lessons from a Startup

While the success of the articles was a major highlight, the most significant takeaways came from the experience of working within a fast-paced startup. This was a proper, intense role that required significant time and commitment.

Because the team was small, they were kind of nurturing out a lot of features quickly. My role involved not just creating content but also testing these new features, providing feedback, and working directly with the co-founders. This was a masterclass in entrepreneurship and product development.

Working so closely with the founders gave me a unique window into their world. I observed how they considered product features, prioritized their roadmap, and addressed the day-to-day challenges of running a business. We even had discussions about funding. This level of exposure was invaluable and completely changed my perspective on how products are built and businesses are run.

My time at Datapane also put me on a new career path. Towards the end of my stint, a product manager joined the company, and I got a couple of weeks to work with him. While I didn't have formal experience, he showed me how product managers think, prioritize features, and work with end-users. This brief but impactful interaction, combined with my experiences as an evangelist, "paved my career into new directions," ultimately setting me on the path to becoming a product manager myself.

The year I spent as a freelance data science evangelist was short but intense. It was a period of rapid learning, unexpected successes, and profound professional growth. It taught me that my skills in data analysis and coding could be used in a completely different way to create value, influence an audience, and ultimately, help a small company achieve its goals.