Population for mini grid planning: finding settled clusters beyond the city
Check that the population grid covers your region, map residents in small hexagons outside town centres, and pick clusters big and close enough for one mini grid.
Part of the guide Location intelligence automation
Using population for mini grid planning comes down to finding clusters: groups of small cells where enough people live close together, far enough from the main grid, to share one generation site and one local network. Map the residents of a rural district in hexagons of about 0.1 km², keep the ones with people in them, group the neighbours, and rank the groups by size. But before any of that, check that the population grid actually covers your country and district. If it does not, every hexagon will read zero and the map will look like empty land.
This is for mini grid developers, rural electrification agencies, solar companies and the non-profits that work with them. It covers the coverage check, how to set up the run for a rural area, how to turn hexagons into candidate sites, the facilities worth adding as anchor customers, and what residents data cannot tell an energy planner.
First, check coverage
xMap's population grid covers about 100 countries, among them India, Thailand, Egypt, France, Canada and Australia, but not every country, and not the United Kingdom or the United States. Coverage of a country is not the same as good detail in every remote valley of it.
- Pick one small area you already know: a market town in your region and a few villages around it.
- Run Demographics on it once, with Totals for each area.
- If the step says the area is not covered, or the count is zero where you know people live, stop: this region needs another source.
- If the total looks right for what you know of the town, switch to every hexagon and look at the map. Villages should show as small warm spots, not one smooth smear.
This takes ten minutes and saves weeks of analysis on data that was never there.
Setting up the rural run
Draw the district with Area of interest. One drawn area is searched in parts of about 170 km², up to 25 parts, so roughly 4,200 km² at most; a bigger district goes in as several shapes, run one after another. Demographics with Give back set to every hexagon returns up to 5,000 hexagons a run, each with its centre, residents, age groups and people per km².
In a rural district most hexagons hold nobody. Add a Filter to keep only those with residents, which leaves the settled land: villages, hamlets and the strips along roads.
From hexagons to candidate clusters
A mini grid serves a cluster, not a hexagon. Neighbouring settled hexagons with few empty ones between them belong together. Two practical ways to group them:
- Put the hexagons on a map with Show on a map as a heat map, draw a shape around each obvious cluster by eye, and run Demographics again with totals for those shapes. Each cluster comes back as one row with its residents and density.
- Ask the Data analyst to group hexagons whose centres are within a set distance of each other and add up the residents of each group. It writes a small program over the hexagons and returns one row per cluster.
Then rank the clusters by residents and by how compact they are. A cluster of 1,500 people spread over 6 km² needs far more cable than the same number in 1 km².
Anchor loads: what is already there
A mini grid pays its way more easily with a few steady daytime customers alongside households. Nearby places, searching within each cluster, finds the schools, health centres, places of worship, shops, mills and workshops listed in xMap's places database. Rural listings can be thinner than city ones, so treat what it finds as a starting list to confirm on the ground, not a full census.
Satellite images of each cluster, about 10 m per pixel, help confirm that the settlement is where the grid says and show its shape: a tight village core or houses scattered along a road.
A made-up example
Say a solar developer is screening a district in eastern India. The coverage check on the district's main town gives a total close to what local officers report, so it goes ahead. It draws the district in three shapes, keeps the settled hexagons, and groups them into 140 clusters. It drops clusters within 2 km of a known grid line, keeps those with at least 800 residents, and searches each for schools and health centres. Twenty-two clusters remain, each with residents, density, and a list of possible anchor customers, ready for field visits.
What residents data cannot tell you
Pairing the clusters with your own data on grid lines and connection records, uploaded to My data, is where the screening gets sharp. For the wider way of putting place and population data to work, see location intelligence automation.
Questions
How do I know if my country is covered?
Run Demographics once on a small area you know well. A zero or a not-covered note where people clearly live means the grid does not cover it.
How big an area can one run handle?
About 4,200 km² per drawn area, searched in up to 25 parts, and up to 5,000 hexagons a run. Larger districts are drawn as several shapes.
Can the data show which villages already have power?
No. It counts residents only. Grid lines and connections need your own data or a field survey.
Are rural schools and clinics all listed?
Not always. Rural listings can be thinner than city ones, so use what Nearby places finds as a list to confirm, not a complete count.
Templates that do this
Read next
More on population and demographics · Every article · Templates