The community, told through data
Two neighbourhoods vote on the same day and behave like two different countries — sometimes on the same street. This atlas is the evidence layer behind that sentence: six data stories built from Elections Canada's own poll-by-poll returns, the City of Toronto's 2021 neighbourhood profiles, the City's own address repository, and a registry of the institutions that actually hold these neighbourhoods together. It maps participation — who gets reached, who gets counted, and where the invitation never arrives — and nothing else. Start with the chart below: it is the finding the stories unfold.
A 20-point participation canyon separates the tower corridor from the lake
This figure described in words (alt text)
Horizontal bar chart of the 15 advance-poll zones of Beaches—East York, ordered from lowest to highest total turnout. The four lowest zones — 608 (East End-Danforth), 603 (Taylor-Massey), 602 (O'Connor-Parkview) and 600 (O'Connor-Parkview/Victoria Village), the riding's tower corridor — run from 55.9% to 59.0%. The six highest — 614, 611, 612, 606, 613 and 609, covering The Beaches, the Woodbine Corridor and Danforth East York — run from 71.4% to 75.7%. The distance from the lowest zone to the highest is 19.8 percentage points, and the low end is a contiguous block of high-rise geography rather than a scatter.Read it this way. The canyon is not apathy and not a mystery: it is tenure, tower geography and language, and it tracks them closely enough to plan around. The four red zones hold roughly 21,000 electors between them. Nothing here identifies a person, a household, or a vote — the smallest unit anywhere on this site is a polling division.
The stories
Two worlds, one street
The same street, two participation regimes. A scrolling map walks the boundary where a 20-point gap opens between neighbours who share a bus stop.
The canyon didn’t close
Four years on, the 20-point gap is still there. E-day turnout moved +1.0 point riding-wide since 2021 and +2.5 in the tower core — neither falling further behind nor catching up.
Tower explorer
Every single-building poll in both ridings, ranked lowest-turnout first, each with the three nearest partner institutions that could open the door.
Pods vs blankets
What full-riding canvass coverage actually costs in human hours — and why the relational model is the only one the arithmetic permits.
Who reaches whom
Three information ecosystems that barely touch, and the population no local channel reaches at all.
The same canyon: Scarborough Southwest
Scarborough Southwest's mirror geography, drawn at the same grain: the zone bars, the tower strip, and the measured street doors. Turnout there is the federal 35096 substitution, stated on every figure.
The data room →
Every dataset behind the stories, sortable and downloadable, with the grain and the caveats stated on the page.
What this is, and what it is not
It is a participation map: turnout by zone, by neighbourhood, by building; the demographic layer that explains it; the measured street doors behind the arithmetic; and the coverage maths that says what a volunteer base can actually reach. It is not a targeting product. There is no candidate data, no party data, no individual-level record, and no persuasion model anywhere in it — by construction, not by policy. Every figure carries its source, its as-of date, and a download link to the same CSV the figure was drawn from, so anyone can check the claim against the data instead of trusting the chart.