FlowsBook a callSign inCreate free accountSign up
Business and places data

How to group businesses into areas for a report

Turn a long list of businesses into a few named areas with counts, ratings and a map. How to choose the areas, tag each business with one, and write the report.

Part of the guide Location intelligence automation

By the xMap team · · 5 min read

To group businesses into areas, decide the areas first (districts you draw, a few named places with a radius, or whole cities), search for the businesses inside each one so every result carries the name of its area, then count and compare by that name. A report built this way reads as "Centre has 120 cafes, the old port 45" instead of one long list, and it can be rebuilt every month without anyone sorting rows by hand.

This article is for analysts, franchise and expansion teams, agencies writing market reports for clients, and anyone who has been asked for "the numbers by area". It covers how to pick areas that make sense, the three ways to tag each business, a made-up example, the traps, and an AI employee that writes the report on a timetable.

Pick areas your reader already uses

A report is only useful if the reader recognises the areas. A city's official districts suit a council or a utility. A retailer may think in trade areas around its stores. A sales manager thinks in territories. Before touching data, write the list of areas and how each is defined:

  • Drawn areas: neighbourhoods or zones drawn on a map, or a boundary pasted in. Best when the shape matters, such as a river or a ring road.
  • Named places with a radius: a station, a mall or a square, each with a circle around it. Quick, and easy to explain: "within 1 km of the main station".
  • Whole cities: when the report compares markets rather than streets.

Keep areas from overlapping if you want counts that add up. If two circles overlap, a business inside both can appear twice, and the total will be higher than the real number of businesses.

Three ways to tag each business with its area

1. Search inside an Area of interest. Put the areas in an Area of interest block, then run Nearby places set to search the Area of interest. Each business comes back with an area field holding the city or drawn area it was found in. One block can hold up to 8 areas this way, each searched from its middle to its edge.

2. Count around a list of centres. If your areas are many points (every station on a line, every one of your stores), keep them in a table in My data and set Nearby places to Each incoming place, with counts. Each centre comes back with how many of the type are within the distance, the nearest one, the average rating and how many belong to a chain. This is the quickest way to a table of areas with a number beside each.

3. Let the territories decide. When you do not have areas yet, Split by territory can cut the businesses into compact groups of about the same size and name them by direction from the middle (North, Centre, South-East). Useful for a first look, or for sharing work, but the names are compass points, not neighbourhoods.

From tagged rows to a report

Once each business carries its area, the report is a matter of counting and saying what stands out. The Data analyst block answers a question such as "how many per area, the average rating per area, and the share of chains" by writing a small program over every row, and returns a table and a chart. An xMap LLM block set to work on the whole list can then write a short summary, and Make a report lays it out as a PDF with the figures, charts by kind and city, a map and the list.

A made-up example in Mexico City

Invented for illustration. A bakery brand is deciding where to open in Mexico City and asks for a monthly page on bakeries in five neighbourhoods it likes. The analyst draws the five neighbourhoods as areas, runs Nearby places for bakeries inside them, and asks the Data analyst for counts, average ratings and the share of independents by area.

Suppose the table shows (made-up figures) 64 bakeries in the first area, 18 in the second, and the second also has the highest average rating. The written summary points out that the second area has few bakeries and the ones it has are well rated, a sign of demand that is not yet served, and that the first area is crowded with chains. The brand's team reads one page and two charts instead of a spreadsheet of 200 names.

Mistakes to avoid

  • Areas of very different size. A district three times larger will have more of everything. Add people or area size where you can, or compare per square kilometre.
  • Overlapping circles. They double count. Draw non-overlapping shapes for counts you intend to add up.
  • Ignoring the search limit. When a search reaches its 500 nearest, the true number may be higher. Smaller areas fix it.
  • Changing the areas each month. A report is read as a trend. Keep the same areas so this month compares with last.
  • Treating a category as exact. Some places are listed under a neighbouring type. Read a sample of each area before trusting the count.

An AI employee that writes it every month

  1. Schedule: on the 1st of every month at 08:00.
  2. Area of interest: your drawn areas, or named places with a radius.
  3. Nearby places: the type of business, Where set to Area of interest, so each result carries its area.
  4. Data analyst: ask for counts, average rating and chains versus independents by area.
  5. xMap LLM, The whole list: three or four sentences on what changed and what stands out.
  6. Make a report: a PDF to your inbox or a Google Drive folder.

On the first of each month you get a PDF with your logo: the summary, the counts by area as a chart, the list of businesses with the area each belongs to, and a map with the areas drawn. Because the areas stay fixed, last month's report and this month's sit side by side and show where the market moved.

Questions

How many areas can one search cover?

Nearby places searches up to 8 areas of an Area of interest in one block. For more, use a table of centres with Each incoming place, with counts, or run the flow for several groups of areas.

Do the areas have to be official districts?

No. You can draw any shape, paste a boundary, or use named places with a radius. What matters is that your reader recognises them.

Why do my area totals add up to more than the businesses I have?

The areas probably overlap, so some businesses are counted in two. Draw areas that do not overlap if the totals must add up.

Can I add how many people live in each area?

Where population data exists, the Demographics block can add residents for each area. It covers some countries, among them France, Canada, India and Australia, but not the United Kingdom or the United States.

Can the report go to a client with my branding?

Yes. Make a report uses your organization's logo and colour, and can email the PDF or save it to a Google Drive folder.

Templates that do this

Read next

More on business and places data · Every article · Templates

Put it to work

An AI employee in xMap Flows does this on live data, on its own timetable. Start from a template or describe it in plain words.