Ian Johnson

A selection of election-related data visualizations created by our community, the Observable Journalism Network, and our team. Explore maps, voter wait times, paths to victory, and more.

Observable

Elections

Some more resources:
- Guide to bivariate choropleth maps: https://www.joshuastevens.net/cartography/make-a-bivariate-choropleth-map/
- Color palette creation guide: https://blog.graphiq.com/finding-the-right-color-palettes-for-data-visualizations-fcd4e707a283

Fabian Iwand

The color palette guide also touches on accessibility.

The importance of square-root based scales for disks: https://observablehq.com/@toja/same-area-different-shapes

Twirls: https://observablehq.com/@bartok32/the-relentless-march-to-musical-oblivion

this is also something really important to consider in the shift maps: https://observablehq.com/@mbostock/methods-of-comparison-compared

added all these to the inspiration gallery, and moved it just below the Table of contents

Data properties:
- size
- color
- stroke width

Variants:
- pie charts
- ellipses (two data axes)

Data properties:
- width
- height
- color
- shape segments

Variants:
- bars
- silos
- arrows

Data properties:
- color
- density (~ opacity)

Variants (abstractions):
- binned
- voronoi
- gradients

Hatching in different directions (horizontal, vertical, diagonal) and densities could be used to display multiple properties.

i love the idea that the angle of the hatching can encode a metric!

Data properties:
- size
- length
- direction
- arrow head/body ratio

Variants:
- straight up, then curved to direction
- sideways triangles
- chevrons
- straight arrows
- https://www.unicode.org/charts/PDF/U2190.pdf

do you think a spike map is just a special case of a shift map?

No, I wouldn't go so far. Spike maps don't incorporate direction, which seems to be the primary aspect of shift maps.

> It might be interesting to try shift map as an alternative/supplement to a bivariate choropleth

For what kind of data? Do you have something specific in mind?

this isn't bivariate but it is showing a shift: https://observablehq.com/@codingwithfire/registered-voters-by-county-2016-2018
could imagine the angle of the line being driven by a bivariate scale and height / color could map to magnitude or something

https://observablehq.com/@clhenrick/health-equity-tracker-data-visualization-precedence looks like another good resource