Quebec City data projectWhich trees are the biggest? Group the measurements first
The Ville de Québec publishes, through the Données Québec portal, an inventory of trees on its territory that it considers potentially remarkable. It is short enough for a beginner and tricky enough to teach a real lesson. All figures below are ours, from the file downloaded on 22 October 2026.
DIAMETRE by measurement position, summarised by us from the Ville de Québec list of potentially remarkable trees (684 trees; downloaded 22 October 2026). The file gives no unit, so none is shown| POSITION_MESURE code | Trees | Median DIAMETRE | Rows with DIAMETRE 0 |
|---|
| DHP | 592 | 79 | 3 |
| M | 65 | 86 | 0 |
| DHS | 27 | 85 | 0 |
A beginner sorts the whole file by DIAMETRE and announces the biggest tree. The file will not allow that so quickly. Each diameter comes with a code saying where on the tree it was measured, and three different codes are in use, so a number from one group is not directly comparable with a number from another. The program below groups the diameters by that code before taking any median, counts the species and checks who owns the land. It uses Python's built-in csv and statistics modules and expects the CSV next to the script:import csv
import statistics
by_position = {}
species = {}
private = 0
with open("vdq-arbrepotentielremarquable.csv", encoding="utf-8-sig") as trees:
for row in csv.DictReader(trees):
by_position.setdefault(row["POSITION_MESURE"], []).append(int(row["DIAMETRE"]))
species[row["NOM_LATIN"]] = species.get(row["NOM_LATIN"], 0) + 1
if row["TYPE_PROP"].startswith("Priv"):
private += 1
for position, values in by_position.items():
print(position, len(values), "median", statistics.median(values), "zeros", values.count(0))
top = max(species, key=species.get)
print(len(species), "species; most common", top, species[top])
print("TYPE_PROP starting Priv:", private)
1. Group before you summarise
Putting every DIAMETRE into one list would mix three measuring methods. A dictionary of lists, keyed by POSITION_MESURE, keeps them apart, and the medians, 79, 86 and 85, are then reported per group, never as one figure.
2. Prefer the median
Three DHP rows record a diameter of 0, which is almost certainly a missing value rather than a tree. A median shrugs off a few such rows; an average would be dragged down. Students then decide, and write down, whether to drop them.
3. Read the other columns
Seventy-five Latin names appear, and Quercus rubra, the red oak, is the most common with 145 trees. TYPE_PROP shows 609 of the 684 trees on land recorded as private, so most of these trees are seen from the street rather than visited.
Further checks on the same Ville de Québec file (our counts)| Check | Result |
|---|
| Distinct tree IDs | 684, one per row |
| Largest DIAMETRE value in the file | 185 |
| Trees recorded with more than one trunk (MULTI_TRONC O) | 62 |
| Rows with a planting date filled in | 8 |
| Trees listed on Chemin Saint-Louis | 168 |
Limits we teach with the data
The file does not say what unit DIAMETRE uses or exactly what each position code means, and we do not guess; a student who needs to know asks the city. A potentially remarkable tree is the city's category, not a ranking, and a list with only 8 planting dates cannot tell us how old the trees are.
The Ville de Québec publishes the dataset on Données Québec with a CC BY 4.0 licence, which allows reuse, including commercial reuse, with credit to the source. Source: Ville de Québec, Arbres potentiellement remarquables (dataset page, licence). The medians and counts are ours, and the city does not endorse this page or our teaching.
