Laval data projectSplash pads, play structures and accents: a Ville de Laval open data project
The Ville de Laval publishes its recreation sites, Sites de loisirs, as open data on Données Québec. The file is in French, column names included. Before any counting, children in our upper classes learn to open it without breaking the accents. All numbers here are ours.
Step 1: the accent trap. The file is saved as UTF-8, where é takes two bytes. Read those bytes with the wrong encoding and every é turns into two strange characters. This four-line demonstration, ours, shows the damage and the repair:word = "Régional"
broken = word.encode("utf-8").decode("latin-1")
print(word, broken)
print(broken.encode("latin-1").decode("utf-8"))It prints Régional Régional and then Régional again. On some Windows computers Python may default to a different encoding, so we teach children to write encoding="utf-8" every time they open a file.
Selected facility types for children, by sector: our count from the Ville de Laval Sites de loisirs file, downloaded 22 October 2026 (1,264 rows)| Sector (nom_secteur) | Modules de jeux (play structures) | Jeu d’eau (splash pad) | Pataugeoire (wading pool) | Butte de glisse (sledding hill) |
|---|
| Secteur 1 | 0 | 1 | 2 | 0 |
| Secteur 2 | 16 | 3 | 4 | 0 |
| Secteur 3 | 12 | 1 | 2 | 0 |
| Secteur 4 | 0 | 0 | 3 | 0 |
| Secteur 5 | 0 | 2 | 3 | 0 |
| Secteur 6 | 0 | 2 | 4 | 1 |
This is our counting script. A set holds the four facility types we want, and a dictionary inside a dictionary keeps the count for each sector:import csv
wanted = {"Modules de jeux", "Jeu d'eau", "Pataugeoire", "Butte de glisse"}
by_sector = {}
with open("sites-de-loisirs.csv", encoding="utf-8") as f:
for row in csv.DictReader(f):
kind = row["descr_type_plateau"]
if kind in wanted:
sector = row["nom_secteur"]
by_sector.setdefault(sector, {})
by_sector[sector][kind] = by_sector[sector].get(kind, 0) + 1
for sector in sorted(by_sector):
print(sector, by_sector[sector])
A label used in two sectors only
All 28 Modules de jeux rows sit in Secteur 2 and Secteur 3. It is very unlikely that four sectors have no play structures at all; more likely they are recorded under another type, or not at all. Students learn to say the file is uneven rather than that the sectors are.
Reading the column names
nom_secteur is the sector name, descr_type_plateau the facility type and ind_sans_reserv a yes or no flag, O or N, which we read as whether booking is needed. That reading is ours; the file does not define it.
Blank rows
Twenty-four rows are blank in every column. The script skips them naturally, because a blank type is not in the wanted set, but a careful student counts them and says so.
What the file cannot tell you
The CSV on Données Québec was last modified in January 2021, so it is a snapshot from several years ago. A splash pad built since then will not appear, and one listed may have changed. Students write the file date beside every result.
Data source: Ville de Laval, Sites de loisirs, Données Québec, licensed under Creative Commons Attribution 4.0 (CC BY 4.0), which allows reuse, including commercial reuse, with credit. Our counts and table are adapted from the original file. No endorsement by the Ville de Laval is implied, and the file’s phone numbers and site names are left out of this page.
