Toronto data lessonWhy the average day on the Toronto Island ferry is not a typical day
The City of Toronto publishes ticket counts for the Island ferries as open data. We took every 2025 row, added up each day ourselves, and found a mean far above the median. The figures are our calculations from the City's file, not numbers the City publishes.
The dataset is called Toronto Island Ferry Ticket Counts, kept by the City's Fleet Services division. Its description says the counts are provided in 15 minute intervals and are updated hourly, and that the ferries operate year-round out of Jack Layton Ferry Terminal. Our download on 22 October 2026 held 284,387 rows, starting on 1 May 2015, each with a timestamp, a redemption count and a sales count. We used only redemptions, which are tickets used to board, and only calendar year 2025.
Daily ferry ticket redemptions in 2025, by month: our sums of the City of Toronto 15-minute rows, downloaded 22 October 2026| Month of 2025 | Days with data | Mean per day | Median per day | Highest day |
|---|
| January | 30 | 571.0 | 586.5 | 829 |
| March | 31 | 877.6 | 838 | 1,518 |
| May | 31 | 3,769.5 | 3,180 | 12,436 |
| June | 30 | 7,992.2 | 5,947.5 | 18,242 |
| July | 31 | 11,784.0 | 10,124 | 24,341 |
| August | 31 | 12,283.3 | 12,020 | 21,160 |
| September | 30 | 5,555.7 | 4,384.5 | 19,598 |
| November | 30 | 983.2 | 874 | 2,288 |
| All of 2025 | 364 | 4,191.9 | 1,690 | 24,341 |
Look at the bottom row. The mean day in 2025 saw 4,191.9 redemptions, yet the median day saw only 1,690. Only 121 of the 364 days were above the mean at all, while 139 days had fewer than 1,000 redemptions. A short run of enormous summer days, 47 of them above 10,000, drags the mean upward. The median ignores how large those days were and simply finds the middle of the sorted list, so it describes an ordinary Toronto day far better.
The short program below reproduces our yearly figures. It reads the file line by line, keeps 2025, adds the 15-minute counts into one total per date, and prints the count of days, the mean, the median and the busiest date. It needs only Python 3 and the CSV in the same folder:import statistics
from collections import defaultdict
day_total = defaultdict(int)
with open("toronto-island-ferry-ticket-counts.csv", encoding="utf-8") as file:
next(file) # the first line holds the column names
for line in file:
_id, stamp, redeemed, sold = line.strip().split(",")
if stamp.startswith("2025-"):
day_total[stamp[:10]] += int(redeemed)
values = sorted(day_total.values())
print("days:", len(values))
print("mean:", round(statistics.mean(values), 1))
print("median:", statistics.median(values))
print("busiest:", max(day_total, key=day_total.get), values[-1])
Quartiles tell the shape
Our lower quartile was 712.25 and our upper quartile 5,885.5. The gap from the median up to the upper quartile is more than four times the gap down to the lower quartile, the numerical signature of a right-skewed distribution.
Weekends in summer
In July 2025 the 23 weekdays averaged 9,881.9 redemptions and the 8 weekend days averaged 17,252.5. A student who reports one July average for the whole month has blended two different populations into one number.
A gap in the record
The file has rows for 364 dates in 2025, not 365: nothing is dated 26 January. Before computing anything, a careful student decides whether that day was quiet, closed or simply missing, and says which assumption the answer depends on.
Licence and attribution
Contains information licensed under the Open Government Licence - Toronto. The file is used under that licence (licence, dataset page); every sum, mean and median on this page is ours. The City of Toronto does not endorse this page or our classes.
