map-distort

Visualise data on a map by distorting the regions to make their areas proportional to the data values

Example

Sample data

Some example data for India is given, including only the states and Delhi. Jammu & Kashmir is taken as the pre-bifurcation state.

population.dat: Population of states in the 2011 census
popincrease.dat: Population increase from 2001-11
LSseats.dat: Lok Sabha seats
gdp.dat: Gross state domestic product (2018) from http://mospi.nic.in
covidcases.dat: Confirmed COVID cases as of 21 May, 2021 from https://www.covid19india.org/
coviddeaths.dat: COVID deaths

First, build the program using

g++-11 -O3 -W -Wall -Wno-unused-result -std=c++2a    weighted_distort.cpp   -o weighted_distort

You would need a modern C++ compiler.

To modify a map you can then use

echo input.map data_input.dat num_rounds | ./weighted_distort > map_output.dat

input.map is the input map with one region per line. Unfortunately, the program only works with regions that are pseudo-svg paths - that is, piecewise linear paths consisting of only "m, h, v, l, z" elements and commas between coordinates replaced by spaces. See states_boundary.map for an example.
data_input.dat contains the data of the weight, one line per region. Regions with zero weight are ignored
num_rounds is the number of rounds to run the region relaxation for. I usually take around 10K rounds, and it stops earlier if every state is resized to 1% accuracy
map_output.dat is the output file which can be easily processed by gnuplot, with x y region_id per line and a blank line between two subregions

Sample images can be generated by running

bash process.sh

This step requires gnuplot and imagemagick (convert) to be callable directly.

The map data for India was created based on Rajesh Odayanchal's map on Wikimedia Commons
Thanks to Samhan Salahuddin for ideas and discussions