Over the last few months, I’ve worked on a few smaller projects that I didn’t think warranted full individual posts on this site, so I just shared them as smaller threads on Twitter directly. But in order to keep this site as a comprehensive collection of the work I’ve done, I’ve decided to make a list compiling this recent work. Some might say this is double dipping on the same content, but my site, my rules! Of course, if you want to see this kind of work first, you should just follow me on Twitter 🙂
* * *
Fixing Georgia’s COVID-19 Dashboard Trickery [July 2020]
After someone pointed out some misleading data visualization practices on Georgia's COVID-19 dashboard, I made a quick comparison of what their dashboard should have showed vs. what it was showing. Georgia was showing the number of COVID-19 cases/tests/deaths by county and color-coding each county from blue to red as the numbers increased. However, they were not keeping these categories consistent, so even if a county got substantially worse over time, the map would not show that change.
Since making the post, Georgia has improved their dashboard's visuals to show their numbers on a continuous gradient with a consistent scale and the ability to compare across time. But this issue was only one of many with this site. Check out the Atlanta Magazine article on the mess that is/was GA's COVID dashboard.
Georgia has been using visual trickery with their COVID-19 dashboard to make their situation not appear as bad as it really is. They are updating the coloring on the map so that things do not appear to change, even though they've gotten much worse. 1/x pic.twitter.com/Yn4j1JiJCn— Ben Tanen (@ben_tanen) July 20, 2020
* * *
Visualizing Musical Genres Using Every Noise at Once [July 2020]
Recently I came across @EveryNoise, which visualizes musical genres in two dimensional space. I’ve always struggled to categorize my own musical tastes so I decided to follow the map set up by Every Noise at Once and see where my Spotify library landed.— Ben Tanen (@ben_tanen) July 26, 2020
Apparently I like rock. pic.twitter.com/U9OUV114Wg
I've always hated the question "what kind of music do you like?" Musical genres are so amorphous and describing musical preferences within these genres is even more complicated and confusing. So when I came across Every Noise at Once, I knew I had to use it to make some kind of guide to describe my own musical taste. The end product was two maps showing definitively what kind of music I like.
Eventually I'd like to make this project into an interactive website so that others can make their own maps, but if you're curious in the meantime, you can find all of the code on Github.
* * *
Revisiting My 2016 Path to Victory Visualization for 2020 [August 2020]
Last year, I put together a visualization that tried to reimagine the way we normally think about the Electoral College and the race to 270. Now that we’re 82 days out from Nov 3, I figured I should revisit this using results from @FiveThirtyEight's newly released 2020 forecast. pic.twitter.com/uaG5c5mzei— Ben Tanen (@ben_tanen) August 13, 2020
* * *
Comparing Biden’s Chances in 2020 Chances to Clinton’s from 2016 (according to FiveThirtyEight) [August 2020]
Finally, following up on the previous post, I was curious how the odds were changing for Biden/Trump in 2020 vs. what we saw in 2016. While Clinton was the favorite throughout most of the 2016 race, there were significant swings in her odds. Thus far, Biden's odds have remained relatively steady, but things can obviously change very quickly, especially as we roll into the final weeks. At the end of the day, these odds may be completely off again anyway, so who really knows whats going to happen...
As of today, most major election forecasts are giving Biden a sizable chance of winning (538 @ 73%, Economist @ 88%), but the same was true for Clinton back in 2016. When we compare 2016 to 2020, it seems like Biden's odds have remained stable, but we've still got time to go... pic.twitter.com/LiNohRWbFh— Ben Tanen (@ben_tanen) August 14, 2020
Checking in on this a few weeks later, the models (538 @ 71%, Economist @ 84%) seems pretty consistent, even through the conventions. We're heading into the final stretch so will see if there are any major mix-ups (October surprise??) pic.twitter.com/i1HUMbA6B0— Ben Tanen (@ben_tanen) September 8, 2020