FlowsBook a callSign inCreate free accountSign up
Business and places data

POI data explained, field by field

What a points of interest record holds, how places are sorted into sectors and types, how fresh the data is, which fields are often empty, and how to put it to work in a flow.

Part of the guide Location intelligence automation

By the xMap team · · 5 min read

POI data is a list of places, each with a name, a type, a position and whatever else is known about it: address, phone, website, rating and number of reviews, whether it belongs to a chain, when it opened, and whether it is still open. Points of interest is simply the mapping term for any place a person might go: a pharmacy, a school, a petrol station, a hotel. The value is in being able to ask "which places of this kind are within this distance of that point?" and get a clean answer.

This guide reads a record the way an analyst would: what each field means, how often it is filled in, how places are classified, how current the data is, and the checks to run before you build anything on it. At the end it shows how to use the data in an xMap Flows AI employee. The location intelligence automation guide puts it in the wider picture.

What one record holds

Here are the fields that come with each place from xMap's places database, grouped by what you would use them for.

FieldWhat it tells youHow to use it
name, categoryThe place and its type, as the database names itFiltering, labelling a map, deduplicating
category_group, sectorThe wider groups the type belongs toRolling up counts: all food outlets, all health services
lat, lngIts positionDistances, maps, drive-time areas, territories
address, city, postal_code, countryWhere it is in wordsMail, routing, matching to your own lists
brand, chainIts brand when it has one, and whether it belongs to a chainSeparating independents from chains
rating, reviewsAverage rating out of 5 and how many people rated itQuality signal; reviews say how established it is
opened, closed, statusWhen it opened, closed if it did, and whether it is openSpotting new businesses and removing dead ones
phone, websiteHow to reach it, when knownSales lists and outreach
traffic_score, avg_stay_minHow busy it is (0 to 100) and how long people stay, when knownRanking partners or sites by activity
inside, floorThe mall or building it is in, and the floor, when knownShopping centre and office tower work

Fields of a place from the Nearby places block. Some are filled in only when known.

How places are classified

Every place sits in a four-level tree. At the top are 12 sectors, such as Food & Beverage, Health & Medical, Retail & Shopping and Travel & Hospitality. Each sector splits into categories (about 50 in all), then groups (about 500), then types (more than 5,000). A coffee shop, for instance, sits under Food & Beverage, then Cafe & Coffee Shop, then the type Coffee Shop.

This matters when you ask a question. "Restaurants" can mean the single type Restaurant, or a whole family of types: fast food, sushi, Mexican and dozens more each have types of their own. Ask for the type alone and you will miss the pizza place labelled as a pizza restaurant. When you pick types in Nearby places you can choose a whole small group in one go, written like "Mexican Restaurant (all 7 types)", and everyday words such as cafes or petrol stations are turned into the right types for you.

How complete and current it is

No places database is a perfect copy of the street. xMap's database holds more than 300 million places and is refreshed every month, and each place says when it was last checked. A business that opened last week may not be in it yet; for those, a live search of online maps is the better tool (the Map business scan block). The other gaps are in the optional fields. Take one real slice:

830Dentists listed
366Rated, of the 460 nearest
4.7Average rating
13Belong to a chain, of the 460
Dentists within 10 km of the centre of Toronto, xMap's places database, November 2026. The rating and chain figures are of the 460 nearest the centre.

So even in a well-covered city, a fifth of those nearest dentists have no rating. Opening dates are similar: of those same nearest dentists, 457 had an opening date and 42 of them opened in the twelve months before November 2026. A filter on rating or opening date silently drops the places without one, which is right for some questions and wrong for others.

Checks before you trust a list

  1. Pull a small area you know well, such as the streets around your office, and compare it with what you see on a walk.
  2. Count how many places have each optional field you plan to rely on. If phone numbers are missing for a third of places, a call list needs another source for them.
  3. Look at the types that came back. If you asked for gyms and see yoga studios, decide whether they count, and adjust the types.
  4. Check closed places. Nearby places leaves out places marked closed unless you switch that off.
  5. Note the date. Write the month on any figure you share, because counts move.

Mistakes people make with points of interest

  • Treating a count as the whole market. One search returns at most the 500 nearest places; when it is full, the step says there may be more.
  • Dividing a subset by the total. If a figure is worked out from the nearest few hundred, keep it to those.
  • Using a radius where travel time matters. Five kilometres across a river is not five kilometres down a high street; pair the data with a Travel area when access matters.
  • Confusing residents with visitors. Places tell you what is there, not how many people live nearby; Demographics answers that, in the countries it covers.

Putting it to work in a flow

In xMap Flows the Nearby places block is the way into this data. It can look around a point, around each of your own places (a My data table of stores, say), or across each city of an Area of interest. It can pass on the places it found, or keep your places and add counts to each: how many competitors within 500 m, the nearest one and how far, the average rating, how many belong to chains.

A typical AI employee runs every week: a Schedule, an Area of interest, Nearby places, Only new, then an email and a map. The template Watch new competitors follows that shape, and Competitor density around each site uses the counting mode. Its notes land in the Team inbox, with the new places listed and the map attached.

Questions

What does POI stand for?

Point of interest: any place a person might travel to, such as a shop, clinic, school, station or hotel.

How often is the data updated?

xMap's places database is refreshed every month, and each place says when it was last checked. For places that opened in the last few days, a live search of online maps finds more.

Why do some places have no rating or opening date?

Optional fields are filled in only when known. Filters on those fields leave such places out, so check how many lack a field before relying on it.

How many types of place are there?

More than 5,000 types, in about 500 groups, about 50 categories and 12 sectors.

Can I count places around my own sites instead of listing them?

Yes. Set Nearby places to pass on each incoming place with counts, and each of your sites goes on with how many places were found around it, the nearest and its distance.

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.