India’s Urban Growth Mapped

About This Chart

Data Visualization Method

• Data: Residential population (GHS-POP R2023A) and built-up surface (GHS-BUILT-S R2023A), both on 1 km grids, combined into a derived index called Reconstructed Urban Intensity on a 0–1 scale. It is not night-time satellite light.
• Period and area: 1975 to 2025 in 5-year steps (11 frames), covering India as defined by the geoBoundaries IND ADM1 file. Lakshadweep lies within the mask but has no cell that passes the display filters.
• Selection: Cells below an intensity of 0.13, or with a built-up share under 1.2%, are hidden. The second rule keeps densely populated farmland from reading as urban. Isolated single cells remain only if they reach a strong core value.
• Sampling: Of the 38,741 cells active in at least one year, 10,800 are shown at the same fixed positions in every frame. They are ranked by peak intensity, long-run growth and the number of years active. No random sampling is used.
• Processing: The 1 km grids are aggregated into 5 km cells using land cells only, so coastal cells are not diluted by sea. Built-up share is the base signal, and population density adjusts it within a limited range. Both sit on a fixed logarithmic scale with the same constants in every year and no per-year rescaling.
• Chart form: An animated timeline map. Each frame shows that year's values only; points are not carried over from earlier years, and cells not yet active stay hidden. Color follows seven fixed classes, from muted blue-teal through amber to ivory, based on intensity rank with a partial boost for cells whose growth since 1975 ranks higher than their current intensity. Point size follows the same value. The landmass is slightly raised to set India apart from its surroundings. There is no pulse, glow or twinkle effect.

Creator's Note

The aim was to place fifty years of India's urban change on one consistent scale, so that a 1975 city core and a 2025 metropolitan region can be read against each other directly. The raised landmass and a restrained seven-color palette keep attention on how settlement spreads, rather than on brightness alone.

In the author's reading, the map does not show a handful of megacities growing on their own. Delhi appears as an established core that widens outward. Bengaluru, Pune and Hyderabad suggest broad growth from smaller 1975 bases. The Gangetic plain carries a wide, diffuse signal, and Kerala appears to rest on an older, dispersed settlement pattern. These are observations of shape, not explanations of cause.

Data Sources

• GHS-POP R2023A — European Commission JRC; residential population per 1 km cell, one input to the index.
• GHS-BUILT-S R2023A — European Commission JRC; built-up surface per 1 km cell, the other input.
• geoBoundaries IND ADM1 — land mask for India, sourced from the DataMeet India community and the Election Commission of India (CC BY 2.5 India); rasterized for this chart.

Limitations

Reconstructed Urban Intensity is an index built for this chart, not an official GHSL product. Its inputs are modeled. GHS-POP distributes census and administrative population onto a grid using built-up information, so each cell is an estimate rather than a head count. GHS-BUILT-S is anchored on satellite collections from a few reference years, with the years in between interpolated.

Uncertainty is higher in the earliest frames, especially for small buildings and scattered settlements. The 5 km aggregation smooths internal city structure, narrow corridors, small towns and islands. The weights, constants and thresholds are the author's choices, and the built-up rule may also remove genuine low-density settlements.

Because only 10,800 of 38,741 active cells are shown, weak rural and small-town signals are under-represented. Color classes are set from India's own distribution, so the same class in another chart does not imply the same absolute intensity. Cell counts reflect display design and should not be used to rank regions.

What This Chart Doesn't Show

The points are not night lights or electricity use. Teal and amber are lower and higher classes of a single index, not two different datasets, and point size is a display value rather than population or area.

The 2025 frame is a GHSL R2023A modeled projection, not a direct observation; the 1975–2020 frames are gridded estimates. The chart does not give urban population growth rates or official metropolitan statistics, and it does not show the causes of change, such as policy, migration or industry. The broad signal across the Gangetic plain should not be read as one continuous city.

The land mask follows the boundary file used in production, which does not separately identify disputed areas. The map is not a statement on sovereignty or effective control.