Cartograms for spatial data visualization Fran cois Libois INRA - PowerPoint PPT Presentation

Cartograms for spatial data visualization Fran¸ cois Libois ∗ ∗ INRA and Paris School of Economics, France Belgian Stata User Group Meeting - September, 6 th 2016 Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 1 / 21

Maps as visualization tools of the world Powerful visualization tools when used with caution... BUT maps may also produce very biased views of the world. Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 2 / 21

Which parts of the world have the lowest population density? Population density by country Population density 369.715 - 24617.92 (30) 194.865 - 369.715 (30) 114.555 - 194.865 (30) 75.88 - 114.555 (30) 51.835 - 75.88 (30) 27.37 - 51.835 (30) 11.26 - 27.37 (30) 0 - 11.26 (30) Cartogram: Mercator projection (conformal) Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 3 / 21

Which parts of the world have the lowest population density? Population density by countries Population density 369.715 - 24617.92 (30) 194.865 - 369.715 (30) 114.555 - 194.865 (30) 75.88 - 114.555 (30) 51.835 - 75.88 (30) 27.37 - 51.835 (30) 11.26 - 27.37 (30) 0 - 11.26 (30) Cartogram: NSIDC-EASE projection (equal area cylindrical) Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 4 / 21

Which parts of the world have the lowest population density? Population density by countries Population density 369.715 - 24617.92 (30) 194.865 - 369.715 (30) 114.555 - 194.865 (30) 75.88 - 114.555 (30) 51.835 - 75.88 (30) 27.37 - 51.835 (30) 11.26 - 27.37 (30) 0 - 11.26 (30) Cartogram: NSIDC-EASE projection (equal area cylindrical) Mercator projection: ◮ larger size distortion as distance to equator increases ◮ inflates the importance of some countries Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 4 / 21

Which parts of the world have the lowest population density? Population density by countries Population density 369.715 - 24617.92 (30) 194.865 - 369.715 (30) 114.555 - 194.865 (30) 75.88 - 114.555 (30) 51.835 - 75.88 (30) 27.37 - 51.835 (30) 11.26 - 27.37 (30) 0 - 11.26 (30) Cartogram: NSIDC-EASE projection (equal area cylindrical) Mercator projection: ◮ larger size distortion as distance to equator increases ◮ inflates the importance of some countries Equal area projection (NSDIC-EASE) ◮ large shape distortion ◮ area preserving Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 4 / 21

Which parts of the world have the lowest population density? Population density by countries Population density 369.715 - 24617.92 (30) 194.865 - 369.715 (30) 114.555 - 194.865 (30) 75.88 - 114.555 (30) 51.835 - 75.88 (30) 27.37 - 51.835 (30) 11.26 - 27.37 (30) 0 - 11.26 (30) Cartogram: NSIDC-EASE projection (equal area cylindrical) Mercator projection: ◮ larger size distortion as distance to equator increases ◮ inflates the importance of some countries Equal area projection (NSDIC-EASE) ◮ large shape distortion ◮ area preserving Are area and shape the relevant characteristics for population density? Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 4 / 21

But most of the population lives in densely populated countries... Population density by countries Population density 369.715 - 24617.92 (30) 194.865 - 369.715 (30) 114.555 - 194.865 (30) 75.88 - 114.555 (30) 51.835 - 75.88 (30) 27.37 - 51.835 (30) 11.26 - 27.37 (30) 0 - 11.26 (30) Cartogram: NSIDC-EASE projection, reweighted by country population using ScapeToad Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 5 / 21

But most of the population lives in densely populated countries... Population density by countries Population density 369.715 - 24617.92 (30) 194.865 - 369.715 (30) 114.555 - 194.865 (30) 75.88 - 114.555 (30) 51.835 - 75.88 (30) 27.37 - 51.835 (30) 11.26 - 27.37 (30) 0 - 11.26 (30) Cartogram: NSIDC-EASE projection, reweighted by country population using ScapeToad The area of countries can represent their relative importance in terms of population Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 5 / 21

The plan... Create a cartogram using ScapeToad (http://scapetoad.choros.ch) Create an (animated) map using Stata and spmap package by Maurizio Pisati Produce a short movie using FFmpeg (http://ffmpeg.org) Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 6 / 21

Cartogram creation: open ScapeToad and add a vector layer Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 7 / 21

Cartogram creation: create cartogram Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 8 / 21

Cartogram creation: choose the spatial coverage Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 9 / 21

Cartogram creation: choose the attribute - the weighting variable Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 10 / 21

Cartogram creation: other layers and constraints Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 11 / 21

Cartogram creation: quality of the cartogram Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 12 / 21

Cartogram creation: computations Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 13 / 21

Cartogram creation: report Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 14 / 21

Cartogram creation: export to shape file Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 15 / 21

Cartogram created Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 16 / 21

French population density Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 17 / 21

Some Stata code Import shape files in Stata using shp2dta by Kevin Crow shp2dta using francecartogram1975.shp, database(deptpop1975) /// coord(deptpop1975coord) replace genid(id) Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 18 / 21

Some Stata code Import shape files in Stata using shp2dta by Kevin Crow shp2dta using francecartogram1975.shp, database(deptpop1975) /// coord(deptpop1975coord) replace genid(id) Create maps use deptpop1975, clear spmap popdens using deptpop1975coord, id(id) /// plotregion(icolor(white)) graphregion(icolor(white)) /// clmethod(custom) clbreaks(0 20 40 60 80 100 150 200 500 1000 25000) fcolor(Rainbow) /// title("France: population by department") subtitle("1975") /// note("Cartogram: Lambert 93 projection, reweighted by d´ epartement" "population using ScapeToad") /// legstyle(1) legtit("Population density") legcount Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 18 / 21

Some Stata code Export the maps and create a video graph export france-popdens1975.png, replace width(960) height(540) winexec "C:/ffmpeg/ static/bin/ffmpeg.exe" /// -report -framerate 2/3 -start number 1975 -i france-popdens%04d.png /// -c:v libx264 -r 24 -pix fmt yuv420p france popdens video.mp4 Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 19 / 21

Some Stata code Export the maps and create a video graph export france-popdens1975.png, replace width(960) height(540) winexec "C:/ffmpeg/ static/bin/ffmpeg.exe" /// -report -framerate 2/3 -start number 1975 -i france-popdens%04d.png /// -c:v libx264 -r 24 -pix fmt yuv420p france popdens video.mp4 Animated slide using beamer in L A T EX \ animategraphics[controls,buttonsize=0.3cm,autoplay,loop, height=0.8 \ textheight] { 0.75 } { "france-popdens" } { 1975 } { 2015 } Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 19 / 21

Many thanks Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 20 / 21

Which parts of the world have the lowest population density? Population density by countries Population density 369.715 - 24617.92 (30) 194.865 - 369.715 (30) 114.555 - 194.865 (30) 75.88 - 114.555 (30) 51.835 - 75.88 (30) 27.37 - 51.835 (30) 11.26 - 27.37 (30) 0 - 11.26 (30) Cartogram: NSIDC-EASE projection, reweighted by country population using spmap Fr. Libois (INRA & PSE) Cartograms 2016 Bruxelles 21 / 21