A visual essay in five beats
The same 15 advance zones stay on screen from electorate to gap to door structure. Then a second riding enters—not as proof, but as the counterargument the first pattern needs.
Focus or tap a Beaches—East York mark to read its zone, neighbourhood, electorate, turnout, and current proxy value.
Beat 1 · One electorate
Each circle is one stable advance zone in Beaches—East York. Circle size follows the number of electors; horizontal order follows only the numeric zone identifier, 600 through 614. Turnout does not determine this opening position.
Beat 2 · Two endpoints
Zone 608 recorded 55.9% and zone 609 recorded 75.7% when election-day and advance votes in this extract are divided by registered electors. The file does not establish travel time or explain the gap.
Beat 3 · Reveal the canyon
All fifteen zones spread between those endpoints. Keeping every mark visible matters: the story is a distribution, not a before-and-after anecdote selected from two convenient places.
Beat 4 · Rearrange by doors
The ratio compares project-recorded L/LE street addresses with unit doors implied by electors ÷ 1.9. Zone 603 sits at 0.05; zone 606 sits at 0.88. The upstream address-point query is not registered, so this is a planning hypothesis—not a verified building classification.
Beat 5 · Let the counterexample in
Thirteen Scarborough Southwest zones enter in grey. Their election-day-plus-advance turnout spans 19.0 points, and similar proxy ratios can sit at different turnout levels. The plot does not control for income, age, tenure, mobility, language, campaign activity, or errors in the proxy itself.
What the visual can support
Under the working assumptions—electors ÷ 1.9 implied doors, then 25 doors per pair-hour—the coverage file totals 1,777 two-hour pair shifts across both ridings. This is scenario arithmetic, not an observed canvass: neither household size nor field speed was measured in these files.
Observed
15 advance zones, one election, one riding, a 19.8-point range.
Suggestive
The 28-zone scatter has a visible tendency, meaningful exceptions, and an incomplete upstream provenance chain.
Not established
No individual behaviour or controlled causal estimate exists in these files.
Election-day-plus-advance votes divided by registered electors for 15 Beaches—East York advance zones, joined to 28 project-derived address-proxy rows across Beaches—East York and Scarborough Southwest.
Advance zones, not polling divisions, are the stable grain. A by-election can renumber polling divisions; zones survive. Merged and void polls carry notes, and receiver polls read slightly high.
A project-derived planning proxy, not an Elections Canada measure or a housing census. Street addresses are rows labelled parcel classes L/LE in the working extract; the upstream address-point query and its retrieval metadata are not registered here. Implied unit doors divide electors by 1.9, and workload divides those implied doors by 25 doors per pair-hour. Both are explicit assumptions, not observed household size or canvassing speed. The ratio cannot identify building type, lobby access, resident motivation, or cause.
The address proxy’s upstream query is not registered. The 1.9 household-size and 25-doors-per-hour values are scenario assumptions. Do not use this file to classify buildings, access barriers, resident motivation, or cause.
Guided argument, then exploration
Sort by riding, turnout, or street-door ratio. The table is complete in the HTML; search and facets only narrow what is already present.
| Riding | Zone | Electors | Recorded street addresses | Implied unit doors (÷1.9) | Address proxy ratio | Election-day + advance turnout |
|---|---|---|---|---|---|---|
| Beaches—East York | 608 | 5,910 | 1,613 | 3,111 | 0.52 | 55.9% |
| Beaches—East York | 603 | 4,948 | 143 | 2,604 | 0.05 | 58.5% |
| Beaches—East York | 602 | 5,178 | 1,180 | 2,725 | 0.43 | 58.7% |
| Beaches—East York | 600 | 4,973 | 1,251 | 2,617 | 0.48 | 59.0% |
| Beaches—East York | 604 | 5,704 | 1,746 | 3,002 | 0.58 | 63.5% |
| Beaches—East York | 610 | 5,679 | 1,924 | 2,989 | 0.64 | 66.1% |
| Beaches—East York | 605 | 5,774 | 2,266 | 3,039 | 0.75 | 66.4% |
| Beaches—East York | 601 | 5,017 | 1,837 | 2,641 | 0.70 | 68.9% |
| Beaches—East York | 607 | 5,392 | 2,065 | 2,838 | 0.73 | 70.7% |
| Beaches—East York | 614 | 4,992 | 1,726 | 2,627 | 0.66 | 71.4% |
| Beaches—East York | 611 | 5,771 | 2,254 | 3,037 | 0.74 | 71.9% |
| Beaches—East York | 612 | 5,539 | 1,798 | 2,915 | 0.62 | 73.4% |
| Beaches—East York | 606 | 5,833 | 2,705 | 3,070 | 0.88 | 73.9% |
| Beaches—East York | 613 | 5,229 | 2,048 | 2,752 | 0.74 | 75.0% |
| Beaches—East York | 609 | 5,393 | 2,199 | 2,838 | 0.77 | 75.7% |
| Scarborough Southwest | 605 | 6,941 | 1,444 | 3,653 | 0.40 | 56.0% |
| Scarborough Southwest | 600 | 6,858 | 926 | 3,609 | 0.26 | 57.9% |
| Scarborough Southwest | 608 | 6,436 | 1,803 | 3,387 | 0.53 | 58.1% |
| Scarborough Southwest | 610 | 6,222 | 1,461 | 3,275 | 0.45 | 58.1% |
| Scarborough Southwest | 601 | 6,328 | 1,572 | 3,331 | 0.47 | 58.1% |
| Scarborough Southwest | 604 | 7,383 | 1,516 | 3,886 | 0.39 | 58.4% |
| Scarborough Southwest | 611 | 7,148 | 2,163 | 3,762 | 0.57 | 58.7% |
| Scarborough Southwest | 606 | 5,980 | 2,337 | 3,147 | 0.74 | 59.2% |
| Scarborough Southwest | 607 | 6,673 | 2,401 | 3,512 | 0.68 | 62.4% |
| Scarborough Southwest | 603 | 6,833 | 2,281 | 3,596 | 0.63 | 62.7% |
| Scarborough Southwest | 609 | 6,690 | 2,162 | 3,521 | 0.61 | 64.0% |
| Scarborough Southwest | 602 | 6,361 | 2,271 | 3,348 | 0.68 | 65.9% |
| Scarborough Southwest | 612 | 6,284 | 2,271 | 3,307 | 0.69 | 75.0% |
28 joined rows. Election-day-plus-advance turnout is from the 45th general election; the address ratio and workload fields are project-derived assumptions documented separately.
Advance zones, not polling divisions, are the stable grain. A by-election can renumber polling divisions; zones survive. Merged and void polls carry notes, and receiver polls read slightly high.
A project-derived planning proxy, not an Elections Canada measure or a housing census. Street addresses are rows labelled parcel classes L/LE in the working extract; the upstream address-point query and its retrieval metadata are not registered here. Implied unit doors divide electors by 1.9, and workload divides those implied doors by 25 doors per pair-hour. Both are explicit assumptions, not observed household size or canvassing speed. The ratio cannot identify building type, lobby access, resident motivation, or cause.
Open the full turnout analysis, including 2021 comparison and advance voting →
