Redrawing the States

This visualization was an attempt by me to:

1. Understand d3 (one day I should really learn it. :) ), and
2. More importantly, to understand just how janky the electoral college is.

Using this visualization, you can move counties to other states. Currently it's a bit difficult to use, but what I found, basically, is that

• If you move the three westernmost counties of the Florida panhandle to Alabama, Florida flips to Clinton,
• If you move the 10 closest counties of the Upper Peninsula of Michigan to Wisconsin, Michigan flips to Clinton,
• If you move the three closest counties of California to Arizona, Arizona flips to Clinton,
• If you move Cook County from Illinois to Indiana, Indiana flips to Clinton (and gains 7 electoral votes), and Illinois flips to Trump (and loses 7 electoral votes),
• If you move Lake County (just above Chicago) to Wisconsin and those 10 counties of the UP to Wisconsin, Clinton wins Illinois, Wisconsin, and Michigan,
• If Camden joined Pennsylvania, Clinton wins both Pennsylvania and New Jersey (and no electoral votes change hands),

In total, if only 8 counties move (3 from CA -> AZ, Camden -> PA, Lake -> WI, 3 from FL -> AL), Clinton wins 301 to 237.

Usage

If you want to try to make sense of the current draft product, then just run

cd public && python3 -m http.server

and then point your browser to localhost:8080/map.html. Or, if you want, go here for the latest live version.

Grabbing data

To grab data and structure it for production, you will need to have both uv and node 12+ installed. After that, you'll need to install dependencies with:

uv sync
npm install

After that, you can create the 2016, 2020, and 2024 data sets by running:

uv run redraw 2016 public/data/us.json
uv run redraw 2020 public/data/us2020.json
uv run redraw 2024 public/data/us2024.json

If you'd like to recreate the 2012, 2008, and 2004 files, you need to grab the data set at::

https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/VOQCHQ

Supposed you saved it as countypres_2000-2016.csv. Then you would run the commands:

uv run redraw mit 2012 countypres_2000-2016.csv public/data/us2012.json
uv run redraw mit 2008 countypres_2000-2016.csv public/data/us2008.json
uv run redraw mit 2004 countypres_2000-2016.csv public/data/us2004.json

Unfortunately, at this time, the Census Bureau's API for 1990 SF1 data seems to be down, and so we cannot create a file for the year 2000. :-/

Acknowledgements

I ganked a lot of stuff from the interwebs to make this. Here is a list:

• Mike Bostock's tutorial on how to make a bubble map underlies a lot of the shape data: link
• Townhall.com's election data by county was used in the original 2016 tool
• In 2020 I moved to the New York Times' data for 2020 and 2016
• Population and income data come from the Census Bureau's decennial SF1 file
• D3 Tooltips from Lee Howorko here
• Colors for the map from FiveThirtyEight's's election coverage
• Lines in the middle of divs from this StackOverflow
• getParameterByName function from this StackOverflow
• The copy-paste examples from clipboard.js are copied verbatim
• Bootstrap, D3, and jQuery are, of course, indispensable
• css-element-queries from @marcj were super useful for zooming in the previous version of this tool
• Connecticut county crosswalk for Connecticut's updated county equivalents in 2022.

Contributors

Kevin Wilson (the owner of the repo) is the main contributor. But some others have helped as well. Notably:

• @herbiemarkwort contributed the "0 population => 0 electors" computation
• @Euonia contributed the keyboard shortcut for going to "Move" mode

License

GPL v3