Burlington data projectWhich Burlington road segments changed most between 2018 and 2022?
The City of Burlington publishes Annual Average Daily Traffic (AADT) counts by road segment, one layer per year, each segment carrying a TES_ID. We downloaded the 2018 and 2022 layers on 23 October 2026 and joined them on TES_ID in Python. Every figure here is our own output.
Largest changes in our join of the City of Burlington AADT layers, 2018 to 2022, matched on TES_ID (our calculation; volumes as the files record them)| Street (2022 file) | 2018 volume | 2022 volume | Change |
|---|
| CAVENDISH DR. | 12,016 | 471 | -11,545 |
| CAVENDISH DR. | 11,038 | 605 | -10,433 |
| GUELPH LINE | 26,578 | 19,123 | -7,455 |
| PLAINS RD. E | 20,147 | 27,603 | +7,456 |
| WALKER'S LINE | 38,350 | 46,439 | +8,089 |
| UPPER MIDDLE RD. | 2,741 | 13,605 | +10,864 |
The join is the lesson. The program first builds a dictionary from the 2018 rows keyed by TES_ID, then looks up each 2022 row in it. A comment in the code marks the trap that the next part of the lesson is about:import json
def load(path):
with open(path, encoding="utf-8") as f:
return [feat["attributes"] for feat in json.load(f)["features"]]
y2018 = load("aadt_2018.json")
y2022 = load("aadt_2022.json")
by_id_2018 = {}
for row in y2018:
by_id_2018[row["TES_ID"]] = row # a repeated TES_ID keeps only its last row
pairs = []
for row in y2022:
old = by_id_2018.get(row["TES_ID"])
if old and row["Volume"]:
pairs.append((row["STREET_NAM"], old["Volume"], row["Volume"]))
print(len(y2018), len(by_id_2018), len(y2022), len(pairs))
pairs.sort(key=lambda p: p[2] - p[1])
for street, before, after in pairs[:3] + pairs[-3:]:
print(street, before, after, after - before)
Keys that repeat
The 2018 layer has 310 rows but 309 distinct TES_ID values, because one segment on Blue Spruce Avenue appears twice. The 2022 layer is worse: 1,687 rows, 1,581 distinct ids, 71 of them repeated.
What that does to the join
The program found 327 matched pairs, but they cover only 294 distinct segments; 52 pairs come from 2022 ids that appear more than once. Counting pairs as if they were segments would overstate the result.
A drop worth doubting
Two Cavendish Drive segments fall from about 11,000 and 12,000 vehicles a day to 471 and 605. The 2018 file labels both Single intersection in its Source column; the 2022 file has no Source column at all, so the program cannot tell us why.
What the table cannot claim
These are differences between two files, not proof that traffic changed by that much. The method behind a figure may differ by year, a segment may have been redrawn, and some 2022 rows repeat. A student who writes “Upper Middle Road traffic quintupled” has trusted the join too far; one who writes what the files record, and what they cannot show, has done data work.
For scale, the busiest 2022 row is on Walker's Line at 48,451 and the median of the 1,675 rows with a volume is 2,964. The City's item description says the AADT data is obtained yearly from Transportation road counts (TES). The files are used under the City of Burlington Terms of Use for Open Data, which grant a worldwide, royalty-free, non-exclusive, revocable licence for any lawful purpose. The live terms server did not respond, so we read them from the Internet Archive copy; the data is from the Navigate Burlington portal. The City has not endorsed this page or our classes, and we have no association with it.
