Twitter published user archives, hooray! I cargo culted some code together off of Mike’s Walmart map to show a little histogram of where I tweet. Two maps with different projections and different visualization of magnitude (size vs color)
Some things I learned:
- Hexbins can be brittle. If I change the scale of the map or the hexes just a bit the qualitative shape of the connected objects changes radically.
- Scaling and centering a projection to a dataset is a little tricky. d3.geo.projection() has you doing the scaling and translation in screen coordinates by default, need to do it in geo coordinates to make more sense. (Mike says to use the center() method.)
- D3 projections are fun! The second projection is a kavrayskiy7, whatever the hell that means, borrowed from this example.