Saint John data projectWhy do so many Saint John house numbers start with 1? Python, math and the city address file
The City of Saint John publishes its civic address points, which it calls "the official source for civic address information for the City of Saint John", as open data under the Open Government Licence - City of Saint John. Our students use them to test a famous piece of math. The counts are ours, from data read on 22 October 2026.
First digit of the 25,401 whole-number civic numbers in the City of Saint John file, against the share Benford's law predicts (our count and our calculation)| First digit | Addresses | Share in Saint John | Benford's law predicts |
|---|
| 1 | 7,970 | 31.4% | 30.1% |
| 2 | 4,765 | 18.8% | 17.6% |
| 3 | 3,013 | 11.9% | 12.5% |
| 4 | 2,334 | 9.2% | 9.7% |
| 5 | 1,974 | 7.8% | 7.9% |
| 6 | 1,645 | 6.5% | 6.7% |
| 7 | 1,260 | 5.0% | 5.8% |
| 8 | 1,276 | 5.0% | 5.1% |
| 9 | 1,164 | 4.6% | 4.6% |
The script behind the table. It uses only Python's standard library, skips anything that is not a whole number, and prints both shares side by side:import csv
import math
counts = {}
skipped = 0
with open("saint-john-civic-addresses.csv", encoding="utf-8") as f:
for row in csv.DictReader(f):
number = row["Civic_Numb"].strip()
if not number.isdigit():
skipped += 1
continue
first = number[0]
counts[first] = counts.get(first, 0) + 1
total = sum(counts.values())
print(total, "whole numbers,", skipped, "skipped")
for digit in "123456789":
share = 100 * counts[digit] / total
benford = 100 * math.log10(1 + 1 / int(digit))
print(digit, counts[digit], round(share, 1), round(benford, 1))
1. Clean before you count
The layer holds 25,577 address points. 176 of the civic numbers are not plain whole numbers, values like 57 1/2, 107 A or 37A, so the script counts and skips them openly rather than crashing or guessing.
2. The math
Benford's law says that in many real collections of numbers the first digit d appears with probability log10(1 + 1/d), about 30.1% for a 1 and 4.6% for a 9. Saint John's civic numbers come close: 31.4% start with 1 and 4.6% with 9.
3. Why it happens here
Numbering starts low on every street and only long streets reach the thousands. In this file 1,947 numbers have one digit, 11,095 two, 10,471 three and 1,887 four, and the largest is 10050. Students explain how that spread pushes the first digit toward 1.
Street designators in the same file, in English and French (our count of address points; the file stores both)| English designator | French designator | Address points |
|---|
| Street | rue | 7,013 |
| Road | chemin | 5,022 |
| Drive | promenade | 3,516 |
| Avenue | avenue | 3,270 |
| Court | cour | 1,590 |
| Crescent | croissant | 1,565 |
| Place and Plaza | place | 410 and 6 |
A translation table that is not one to one
Leaving out blanks, the file has 48 different English designators but only 47 French ones, because Place and Plaza both become place. A student who builds an English-to-French dictionary from it works fine; one who builds the reverse dictionary finds that place now means Place, and the 6 Plaza points have silently changed type. Seven rows have no designator in either language. Finding that kind of quiet loss is what data work in a bilingual city looks like.
We never print individual addresses; the project uses only counts. Contains information licensed under the Open Government Licence - City of Saint John (open data catalogue, dataset and licence text). The licence carries no endorsement of our classes, and the City gives none.
