Hotel Building Polygons (Footprints)

For each hotel, the footprint of the building it occupies, as a WKT MULTIPOLYGON in WGS84 (EPSG:4326), with its area in square metres, where the footprint came from, and a grade from A to E for how confidently the building matches the hotel.

Fields

Polygon sources

Match grades in the sample

215 of the 250 sample hotels (86%) have a polygon; the median footprint is 946 m². Grades are match confidence: A is the highest confidence and E the lowest. We recommend grades A–B for automated use and a review for C–E.

Footprints from the sample

The WKT of the first footprint, as it appears in building_polygon_wkt:

MULTIPOLYGON (((2.6967423 39.7994039, 2.6968025 39.7992372, 2.6970276 39.7992449, 2.6970236 39.7993252, 2.6971609 39.7993453, 2.6971311 39.799438, 2.6971048 39.7994342, 2.6971057 39.7994244, 2.6969995 39.7994209, 2.6969874 39.7996061, 2.6967543 39.7995984, 2.6967637 39.799406, 2.6967423 39.7994039)))

How to load it

The polygons are plain WKT text in a CSV or Excel column, so any GIS tool reads them. The guide Load hotel polygons (WKT) covers each tool step by step.

GeoPandas

import pandas as pd, geopandas as gpd
from shapely import wkt

df = pd.read_csv("hotels.csv")
df = df[df["building_polygon_wkt"].notna()]
gdf = gpd.GeoDataFrame(df, geometry=df["building_polygon_wkt"].apply(wkt.loads), crs="EPSG:4326")

PostGIS

ALTER TABLE hotels ADD COLUMN geom geometry(MultiPolygon, 4326);
UPDATE hotels SET geom = ST_GeomFromText(building_polygon_wkt, 4326)
WHERE building_polygon_wkt IS NOT NULL;

QGIS

Layer → Add Layer → Add Delimited Text Layer, choose the CSV, then under Geometry definition select Well known text (WKT), pick building_polygon_wkt as the geometry field and EPSG:4326 as the CRS.

Converting to GeoJSON

gdf.to_file("hotel_buildings.geojson", driver="GeoJSON")

Or with GDAL: ogr2ogr -f GeoJSON hotel_buildings.geojson hotels.csv -oo GEOM_POSSIBLE_NAMES=building_polygon_wkt -a_srs EPSG:4326

Uses

• Geofencing and footfall attribution: a building outline is a tighter fence than a radius around a point.
• Property size estimates: building_area_m2 gives the footprint area of each hotel.
• Mapping and GIS: draw hotels as buildings, not dots.
• Insurance and real estate: link a hotel to its building and to the other places inside it (places_inside_building).

See mapping, GIS and geofencing.

Attribution

Building footprints are derived from OpenStreetMap (© OpenStreetMap contributors, ODbL) and Overture Maps Foundation data, and keep their original licences. Details are on the methodology page.