May 13th, 2021
Hi everyone! 🙋🏻♀️ Elana here. This week Lisa’s off in Mongolia giving a data visualization training, and asked me to stand in for her, so in her honour I’ve prepared something inspired by one of her Datawrapper training wisdoms.
Lisa once mentioned that she often advises her students that “the scatter plot is like the Photoshop of Datawrapper”.
That stuck with me. Because she’s right, it really is such a flexible and powerful chart type that with a little creativity and patience, it’s basically a canvas that you can turn into just about anything.
Well, I decided to take this one step further
away from dataviz. After all – what is Photoshop if not a way to define the specific position of colored points in two dimensional space? And isn’t that kind of what a scatterplot is too?
How hard would it be to paint a picture using our scatterplots, I wondered to myself.
Well, by digging around in my toolkit not that hard as it turns out. So how did I do this? If you’re interested, read on. If you just want to play with the final result, hover over the chart and click Edit this chart.
First I had to find an image, then I opened it in Photoshop and reduced the size to
65px x 78px
Then I used Photoshop’s posterize to reduce the number of colors.
I used Processing to write a little script to go through all of the 5078 pixels of the reduced image, extracting the rgb color values and coordinates of each pixel, feeding this data into a csv file.
I then imported my csv file into Datawrapper. And out popped… a big blue square.
Not quite what I was after! I still had to customize the colors. I could have done this by hand, but with 133 different color groups, I decided to use Datawrapper’s API instead.
A little further adjustment of my axes, and then I was pretty much done. But I wasn’t quite satisfied. What about the tooltip? Well, I decided to get a little creative there too, and with a bit of CSS, I was able to visualize the percentages of red, green and blue for each pixel directly within the tooltip.
And that’s it! Hover over the chart and click Edit this chart if you want play with the final result directly in Datawrapper.
Hope you guys enjoyed this not so charty weekly chart! Stay tuned for next week when Lisa will be back, with possibly fewer Baroque paintings, and a little bit more data!