Geospatial automation without code: the GIS jobs you can now schedule
Buffers, drive times, nearest neighbour, point in polygon, territories and routes: which map operations geospatial automation can repeat for you, and how each maps to a block.
Part of the guide Location intelligence automation
Geospatial automation means running the standard map operations (a buffer around a site, a drive-time area, the nearest depot to each customer, which points fall inside a zone, balanced sales territories, a visiting order) on a timetable or when new data arrives, instead of redoing them by hand in desktop GIS software. Most of these operations are well understood and repetitive, which makes them good to hand over to a workflow that needs no code.
This article is for analysts and operations people who know what a buffer or a spatial join is, and for managers who have only heard the terms. It translates each common operation into the plain job it does, shows which no-code block performs it, works through a made-up example in Melbourne, and lists the mistakes that make automated map results wrong.
Six operations, in plain words
| GIS term | The plain job | Block in xMap Flows |
|---|---|---|
| Buffer | Everything within a set distance of a point | Area of interest with a radius, or Nearby places within a distance |
| Isochrone (drive-time area) | Everything reachable in N minutes by car, on foot or by bike | Travel area |
| Nearest neighbour | For each place, the closest of my own sites | Find nearest |
| Point in polygon | Did this thing enter, cross or leave a zone? | Zone alert |
| Balanced partition | Split places into fair territories by area | Split by territory |
| Travelling salesman | The best order to visit a day's stops | Route planner |
Common geospatial operations and the blocks that run them.
Two more are worth knowing. A spatial join (attach the attributes of whatever is around a place) is what Nearby places does when it passes on each of your places with counts of what was found around it. Aggregating to a grid, such as hexagons, is what Demographics does when it returns residents per hexagon for a heat map, in the countries where population figures exist.
Why automate them at all
Each operation is quick once. The work is in doing it again: when the list of stores changes, when a rival opens, when a new month's data arrives, when a new rep joins and territories need redrawing. In a manual workflow each repeat is an export, a reprocessing step and a new file, and the file is out of date the following week. Automating the chain means the inputs (your sites, an area, a data source) are defined once and the outputs refresh themselves.
It also keeps the method the same. When a buffer is set to 1 km in a saved flow, nobody quietly uses 800 m next quarter. Comparisons over time stay fair.
A worked example, made up
A made-up chain of 14 physiotherapy clinics in Melbourne wants a monthly picture of its network. Done by hand, it took an analyst most of a day. As an automated chain it looks like this:
- Start from the clinics' own table, each with its position.
- Draw a 15-minute drive area around each clinic (the isochrone).
- Count physiotherapy and sports medicine clinics inside each area (the spatial join), with how many belong to a chain.
- For each rival found, find the nearest of the chain's clinics and its distance (nearest neighbour), to see which clinic faces each rival.
- Add residents per drive area, since Australia has population figures.
- Write a one-page report per month comparing each clinic's area with the month before.
In the first report, two of the clinics share most of their drive area, which the owners had suspected but never measured. One suburban clinic has no rival inside its 15 minutes at all. The figures are invented here, but this is the kind of finding that only appears when the same operations run on every site in the same way.
Mistakes that make automated map results wrong
- Circles where drive times belong. A 3 km circle across a river or a motorway includes places nobody would travel to. Use a travel area when the question is about reach.
- Positions you never checked. If a site's coordinates point at the middle of the city, every buffer and count around it is wrong. Check your own table's positions on a map once.
- Counting the same place twice. When two areas overlap, a rival inside both is counted for each. That is right for "rivals per clinic" and wrong for "rivals in the network"; remove duplicates for the second.
- Comparing across countries without checking coverage. Population grids and some road data do not exist everywhere; a missing figure is not a zero.
- Forgetting the estimate flag. When a drive-time area cannot be drawn on the roads, a circle at a typical speed is used instead and said to be an estimate. Read the note before quoting the result.
Building the chain as an AI employee
The clinic chain's monthly job, block by block, in xMap Flows:
- Schedule: the first Monday of each month.
- My data: the clinics table.
- Travel area: 15 minutes by car around each clinic.
- Nearby places: the rival kinds of place inside each area, passing on each clinic with its counts.
- Demographics: residents within each area.
- Calculated field: residents per rival.
- Make a report: the monthly PDF with tiles, a map and the table, emailed to the owners.
Other chains use the same building blocks differently: Find nearest for the closest depot to each delivery, Zone alert for ships entering a port area, Split by territory and Route planner for field teams. Ready-made starts include Competitor density around each site and Split my city into fair sales territories. For the concepts, see drive time vs radius and geofencing; for more jobs, the guide to location intelligence automation.
Questions
Is geospatial automation the same as GIS?
It uses the same operations, but runs them as a saved chain on a timetable instead of by hand in a desktop tool. Deep one-off analysis still suits GIS software; repeated jobs suit automation.
Do I need to know coordinate systems or file formats?
No. Places carry their latitude and longitude, areas can be picked by name or drawn on a map, and the blocks handle the rest. You can still paste GeoJSON if you have it.
How large a drive-time area can be drawn?
It depends on the plan the travel areas are drawn on; the block's choices stop at what is allowed. Larger reach can be approximated with a distance by road.
Can the results go back into my own systems?
Yes. The output can be written to Google Sheets, a table in My data or your own database, as well as emailed as a report.
Templates that do this
Read next
More on business and places data · Every article · Templates