200,000 words later

Hey there, it’s Fabian. I’m a developer at Datawrapper and work on chart publication features and some new UI designs.

2 years ago on June 31, 2018, Lisa published a blog post called “Your Friendly Guide to Colors in Data Visualisation”. At that time, I wasn’t working at Datawrapper but was looking for inspiration and guides on how to create good data visualizations. I read Lisa’s post and from then on I followed Datawrappers Blog “Chartable”.

In the next months, I read almost all of the new posts, learned a lot, and tried to create my own visualizations with Datawrapper. In October 2018, another post got published: “We’re hiring a software engineer in Berlin”. Datawrapper was looking for a software engineer, the role sounded really interesting and since you’re reading my post now on the Datawrapper blog, you know what happened next.

Today’s Weekly Chart is about our blog from its beginnings in November 2012 until today. Since then, 23 amazing people have written 290 posts and 203,605 words (excluding this one). As a comparison, “Harry Potter and the Deathly Hallows” is 198,227 words long (source).

It all started with a “Hello world, hello Datawrapper 1.0” from Gregor and continued with feature announcements and a couple of user interviews. Some years pass and the blog keeps growing. Then, in October 2017, something big happens.

Lisa starts to work at Datawrapper – and with her, the number of blog posts exploded. All categories got more posts and Lisa introduced the first Weekly Chart: “I’m surrounded by old people”. This was the first of 145 Weekly Charts. Since then, not a single week has been without one (and we are going to keep it that way).

It was a lot of fun to take a closer look at the Datawrapper blog and I encourage you to check out some of these posts. Tap, hover, or click on the bubbles in the first visualization to get more information about every individual post.

How I created the Visualization

The Datawrapper blog is statically generated with a tool called Hexo, and all the code and content sits in a repository on Github. Since all posts are part of the code and written in Markdown, I could write a script that iterates over them and extracts metadata like title, authors, category, and some more. We save that information in a format called “Front Matter”. It looks like this:

title: "200,000 Words Later"
permalink: weekly-chart-analyzing-the-datawrapper-blog
author: Fabian Gündel
categories: Weekly Chart
description: "We took a deep look at our blog and analyzed all posts over the last 8 years."

I wrote the extracted data into a SQLite database to run SQL queries and get a feeling of what information might be interesting. I quickly decided to group the posts by categories and showed them over time:

This first iteration looked a little odd and it was very hard to hover over bubbles to get more information. That’s when Lisa suggested using “jitter” and spread the bubbles out a little more. All post categories were encoded with numbers from 1 to 3, which made adding jitter very easy.

If you are interested in more in-depth information, check out the Weekly Chart “Jittering, explained”.

That’s it for today! I hope you enjoyed this little story. Next week’s Weekly Chart will come in a week – that’s for sure. See you then!