Open-data research

India District Growth Atlas

An explorer and visual studies of India's built environment, population, and economic activity.

Built space and economic output

Building volume per person and nominal GSDP per person. Each dot is a state or union territory.

Log-scale comparison of building volume and GSDP per person across 33 states and union territories. Pearson correlation +0.52. Delhi and Chandigarh are hollow markers. — 33 states / UTs · log–log Pearson r = +0.522023 volume and population estimates; FY2022–23 nominal GSDP. The dashed line is a log–log fit. Axes stay fixed between views.

The relationship strengthens when Delhi and Chandigarh are removed. Both measures share a population denominator; this is a descriptive comparison.

The district view

Modeled building volume across 640 districts, using the source’s 2011 district geography.

Map of total modeled building volume across 640 Indian districts in 2023. A logarithmic blue scale shows total volume, rather than volume per person or per unit of land. — Total volume, rather than density. Building volume is a satellite-derived measure, not usable floor area.

Growth tells a different story

Nominal GSDP growth compared with nighttime-light growth across the same 33 states and union territories.

Scatter plot of nominal GSDP growth from FY2015–16 to FY2022–23 against nighttime-light growth from 2016 to 2023. Pearson r is -0.04, with little linear association in this sample. — 33 states / UTs · Pearson r = -0.04GSDP: FY2015–16 to FY2022–23. Nighttime lights: calendar years 2016–23. Dashed line: linear fit.

In this sample, growth in lights is a poor stand-in for growth in nominal economic output.

Data and method

I built the explorer, aggregated district data to states and union territories, and compared levels and growth. The figures here replot the original analysis inputs.

Satellite data: XKDR Forum, using Google Open Buildings and VIIRS. Geometry and measures are joined by district ID. Andhra Pradesh / Telangana and Jammu & Kashmir / Ladakh follow the harmonization in the original scripts.

GSDP and population estimates are the values recorded in that analysis, rather than newly verified official estimates. GSDP is nominal; volume is modeled. The population-adjusted partial correlation of log total volume and log total GSDP is +0.40 for all 33 states / UTs, and +0.66 after excluding Delhi and Chandigarh. That adjustment does not establish causation.

Comparison data ↓Recorded inputs ↗Calculation notes ↗Satellite source code ↗