Regina data projectTwo City of Regina layers, two answers about playgrounds
The City of Regina publishes its parks as open data in two layers: park areas with yes or blank columns for each amenity, and a separate set of amenity points. Both describe playgrounds. Our students check whether they agree. The counts are ours.
Our reconciliation of the City of Regina ParksAndAmenities layers, downloaded 22 October 2026 (998 park areas, 977 amenity points), matched on the NAME column| Measure | Our count |
|---|
| Park areas in layer 0 | 998, under 442 different names |
| Park areas with PLAYGROUND marked Yes | 210, under 200 names |
| Playground points in layer 1 | 254, under 220 names |
| Names in both lists | 175 |
| Names flagged Yes but with no playground point | 25 |
| Names with a playground point but not flagged | 45, of which 35 do not appear in layer 0 at all |
Here is the script. It uses only the json module in Python 3, with both downloaded files in the same folder, and Python sets do the matching:import json
with open("parks-layer0.geojson", encoding="utf-8") as f:
parks = [feat["properties"] for feat in json.load(f)["features"]]
with open("amenities-layer1.json", encoding="utf-8") as f:
points = [feat["attributes"] for feat in json.load(f)["features"]]
flagged = {p["NAME"] for p in parks if p["PLAYGROUND"] == "Yes"}
pointed = {a["NAME"] for a in points if a["ASSET_TYPE"] == "PLAYGROUND"}
print("park areas:", len(parks), "distinct names:", len({p["NAME"] for p in parks}))
print("playground points:", sum(1 for a in points if a["ASSET_TYPE"] == "PLAYGROUND"))
print("names flagged Yes:", len(flagged))
print("names with a point:", len(pointed))
print("in both:", len(flagged & pointed))
print("flag only:", len(flagged - pointed))
print("point only:", len(pointed - flagged))
One park, many shapes
The first surprise is that 998 areas carry only 442 names. A large park can be drawn as many polygons, so counting rows counts pieces, not parks. Turning names into a set fixes that before any comparison.
Set operations
With two sets of names, the questions become one line each: the intersection gives names on both lists, and the two differences give names on only one. A Grade 9 student can follow it, and it is exactly the kind of data structure work Computer Science 30 names.
Why the key fails
Of the 45 point-only names, 35 never appear among the park areas. That points to names spelled or recorded differently in the two layers rather than missing playgrounds. A text name is a weak key; the stronger next step is a spatial test of which area each point falls inside.
What a careful student writes
Not “the city has 254 playgrounds”, and not “45 playgrounds are missing from the map”. A fair sentence is that the two layers agree on 175 park names and differ on 70, and that most of the differences look like naming mismatches. The file supports that, and nothing stronger.
Regina’s winters show up in the same file: the amenity layer lists 34 outdoor rink shelters, 31 unboarded and 24 boarded outdoor rinks, 19 toboggan hills and 7 cross-country ski trails, a good second exercise in grouping categories. Contains information licensed under the Open Government Licence - City of Regina (licence, ParksAndAmenities service). No endorsement by the City of Regina is implied.
