Cloudstitch launched a new product called D3 in a Box this week, and I wanted to write a few notes about the motivation behind it. It’s one of those “what we found is not what we expected” stories.
D3 in a Box is like a docker container for a D3 visualization: it bundles up everything you need to make and manage a production-quality graphic. Just keep your data in a Google Spreadsheet and manage your code Cloudstitch’s cloud-based editor, and we take care of all the magic to deploy it, scale it, and clone it as many times as you like.
In fact, the best way to get started is to clone an existing visualization and modify it. Here are some examples from our starter-project gallery:
Just click clone and you get your own container filled with everything you need for a production-scale visualization.
When we started building D3 in a Box, we didn’t realize project management and deployment was a useful feature for visualizations. Our original idea was to adapt D3 visualizations so that non-coders could customize everything about them with a spreadsheet.
We won a Prototype Grant from the Knight Foundation to work on spreadsheet-powered news apps. We built a number of evergreen visualizations, like electoral maps and sports brackets. And we enabled all sorts of elaborate spreadsheet-based control panels to modify their look and feel. Here’s just one worksheet of settings from a map widget:
D3 in a Box was initially aimed at making D3 widgets hyper-customizable with spreadsheets.
But the more we visited newsrooms and chatted with developers, the more we realized this was missing the mark. All the visualizations we encountered fell into roughly one of two categories. Standard go-tos like maps and scatterplots all had slick custom products built up around them, like MapBox or Tableau, so why use a spreadsheet? The other half were so custom and story-specific that programming, not pre-set spreadsheet options, was the better level of abstraction.
But while we were getting schooled in newsroom programming, we noticed something that was in dire need of improvement: visualization workflow involves a lot of overhead. You’ve got to manage data gathering, archive code, support reuse, train non-coders, host the darn thing, and often wrestle with an uncooperative CMS. One newsroom developer showed me all the custom apps he had built just to crowdsource data collection across the newsroom. Another showed me a computer under the desk with a post-it note that said “DO NOT TURN OFF.” It hosted all the paper’s visualizations separately from machine running the CMS. And all this is time that’s not being spent on the actual story.
So that’s what D3 in a Box fixes: all the overhead work you’d rather ignore so you can get back to telling your visual story.
Just like the Docker folks realized that there’s great value in packaging and encapsulating server configurations, we think that makes a lot of sense for visualizations. D3 in a Box visualizations are self-contained, full-stack bundles that just work. If you’ve been spending time writing data collection app, just mail out a link your D3 Box’s Google Sheet instead. If you’re hosting visualizations on a dusty machine under your desk, we’ll take the baton for you. If your CMS is a grump, we integrate with it anyway.
Here’s a short video tour of cloning and editing a Box:
Video tour of cloning and editing a D3 in a Box visualization.
Let us know what** **you think about this model — we’d really like to be of service to the all the newsrooms out there. Feel free to leave a comment on this post or shoot me a note at email@example.com.
By clapping more or less, you can signal to us which stories really stand out.