Ways to Improve a Map Visualization

How long does it take you to decode the first picture vs the second?Visualization: Shelby Temple, made with Tableau; data: Tableau Sample SuperstoreVisualization: Shelby Temple, made with Tableau; data: Tableau Sample SuperstoreThere are many more reasons to bring geography into your data visualization but I feel that this is enough to agree that this type of visual has significant advantages..As promised, let’s jump into some tips on how to take your map visualization to the next level.1..Use Viridis or Magma Color PalettesWhen selecting colors for data visualization I’ve noticed people are often selecting the colors that appeal the most to them..Unfortunately, this is not the best approach — because what appeals to you might not be appealing to someone else.A data visualization needs to have colors that will best resonate with the audience..In addition, it should lead to quick and accurate decisions..Using certain color palettes (such as Jet, the old default for Python package Matplotlib) were found to lead to slower and inaccurate decision making.No need to worry, Nathaniel Smith and Stefan van der Walt have already created open source palettes that are colorful, pretty, accessible to colorblind viewers, and now research backed to be effective color palettes for quick and accurate decision making.Visualization: Shelby Temple, made with Tableau; Inspired by work of Jacob Olsufka; Data: Temperature data from MarkieAnn PackerJet suffers from not being perceptually uniform or color blind friendly..Heat under indexes on range of colors and washes out easily.If you want to learn more about color theory and the development of Viridis this video from the creators is both entertaining and informative.Also, here is an article where Viridis was tested against other common color palettes and performed well for both speed and decision accuracy.A quick thank you to Ben Jones who’s Tweet originally brought these color palettes to my attention.2..Use Hex Tile Grid MapsHex tile grid maps were recently created as an improvement to the tile grid map..Tile grid maps were created to deal with issues such as trying to create a map of the US with Alaska and Rhode Island..Where Alaska demands too much visual real estate and Rhode Island often gets lost..Here I am plotting out average tornado track length by state.Visualization: Shelby Temple, made with Tableau; data: https://www.ncdc.noaa.gov/stormevents/Visualization: Shelby Temple, made with Tableau; Inspired by work of Matt Chambers; data: https://www.ncdc.noaa.gov/stormevents/These maps are much like the United States’ Senate, where every region gets equal (visual) representation..Rhode Island is now the same size as Alaska, but the positioning still allows the audience to understand they are looking at a map visualization of the US.The hex grid map improves on the tile map by allowing for more adjacent shapes that are the same distance away.. More details

Leave a Reply