Ontario Capacity Explorer

Scale

Context

Fiscal years

to

Data

Ontario capacity

What this is

Ontario’s public capacity, measured against the number of people it has to serve. Hospital beds, physicians, long‑term care, school staffing and provincial jails — counted per person rather than in totals, because a province that adds capacity more slowly than it adds people is losing ground in years when nothing visibly closes. Raw counts are here too, and the same concept is offered under both, so you can see a number fall while the count behind it holds steady.

Every figure is transcribed from a published source and carries that source with it, down to the year. Where a source has no figure, the gap stays a gap: nothing here is interpolated, smoothed or estimated to make a line continuous, and a missing year is never drawn as a zero. Where two sources have been joined into one line, they overlap somewhere and agree there — and the build re-checks that agreement every time it runs.

Status

pre‑alpha Not yet alpha. The series list is still changing, the history is still being pushed backwards, and figures may be corrected or withdrawn as sources are checked against each other. Treat anything here as provisional and check it against the cited source before relying on it.

Changelog

16 September 2026

Data

Ontario’s population now reaches back to 1921, thirty years further than before, from Statistics Canada’s historical provincial estimates. The province held 2,933,700 people in 1921 and 4,471,000 by 1950. The age detail comes with it: Ontarians aged 75 and over went from 54,500 to 127,500 across those years, and children aged 5 to 17 from 738,600 to 886,400. Those two lines gain more than the total does: the July 1 estimates carry no age detail before 1971, so the twenty years from 1951 to 1970 were a hole in the middle of both and are now filled.
The early figures are June 1 where everything from 1951 is July 1, because Statistics Canada publishes no July 1 provincial estimate reaching further back. They join the same line rather than sitting beside it: the two bases are about 0.26% apart — one month of growth — so the step from 1950 to 1951 is thirteen months of it rather than twelve. The reference date is not assumed: in all six census years covered the estimate is exactly the census count, and those censuses were taken on June 1. Each row carries its own date, which the table view shows in its As-of column, and the build re-runs both the census test and the size of the offset on every run.
Enrolment is one line again, from 1866‑67 to today. The twenty‑one school years that used to sit between the Confederation‑era series and the modern one are now eighteen years of data and a three‑year hole. Ontario public elementary and secondary enrolment for 1975‑76 to 1993‑94 was read out of Statistics Canada’s Education in Canada (cat. 81‑229), whose provincial table prints public, private, federal and blind‑and‑deaf schools beside the total of the four. The volumes are scans whose text layer drops the row labels, so every year was reconciled against that printed total and any that failed was discarded; all but four of the nineteen are corroborated by a second edition as well. Enrolment fell from 1,994,638 in 1975‑76 to a floor of 1,759,451 in 1984‑85, then climbed to 2,039,709 by 1993‑94.
That is also what licenses the single line. Section W and the volumes both publish 1975‑76 — 1,995,000 against 1,994,638 — and Section W rounds to the nearest thousand, so the volume’s figure rounds to Section W’s exactly. The build checks it on every run, as it already does for teachers.
The join at the other end is not verified, and the line says so by breaking. The 1995 edition is the last to carry the table, and the modern survey begins at 1997‑98, so 1994‑95 to 1996‑97 have no source at all — no table in Statistics Canada’s catalogue covers them. Those three years are loaded as “no data” rather than left out, so the chart lifts the pen instead of drawing a straight line across a hole. What can be checked is that nothing jumps: the growth implied across the gap is 0.68% a year, against 1.81% in the five years before and 0.65% in the five after.
Worth knowing when reading the 1980s: public enrolment turns upward in 1985‑86 while private enrolment falls by a quarter. Ontario extended full public funding to Catholic secondary schools, which moved students between the two columns rather than into the province.
Each series now records what it is counted in as well as what it is called, so a chart can tell which lines are the same kind of quantity. Educators, students, teachers, physicians and Ontarians are all counts of people, and a selection of them now plots on one honest scale instead of being broken into a panel each. The axis is labelled with what they are counted as while they agree — “educators” — and with “people” once they don’t. Nothing else was merged: a bed is not a person, a prison space is not a bed, and every rate keeps its own denominator, so those still take an axis each. One scale is not always the readable one — a quarter of a million educators against fifteen million Ontarians is a flat line along the bottom, which is what “Indexed” and “Two axes” are there for.
The teachers line now runs to 1993‑94, and it is teachers again. Full‑time teachers in Ontario public schools were transcribed for 1975‑76 to 1993‑94 from Statistics Canada’s Education in Canada (cat. 81‑229), whose annual volumes print the provincial table and reprint each school year two or three times over as later editions revise it. That joins the Confederation‑era Section W series rather than sitting beside it, because in both years the two sources report a final figure they publish the same Ontario number exactly — 95,100 and 94,156 — which is a closer overlap than anything else spliced here. The build checks it on every run. Ontario had about 96,500 teachers in 1975‑76, fell to about 91,400 by 1982‑83, and climbed to 121,605 by 1993‑94.
That is where teachers stop. The 1995 edition is the last to carry the table, and the modern survey publishes educators — teachers plus school administrators plus pedagogical support — from 2002‑03 and never teachers alone. The eight school years between them have no staffing figure at all. Both holes are now written down on the Sources & gaps tab, along with why the teacher/support‑staff split cannot be had: no table in Statistics Canada’s catalogue breaks educators into their three categories, educational assistants are excluded from the count outright, and Ontario’s own school board staff dataset is unreleased.
The educator line says what it contains. Its description now spells out that it counts administrators and pedagogical support beside classroom teachers, and its notes carry two reporting changes that are invisible in the line itself: early childhood educators were added in 2010‑11, and supply teachers were dropped in 2013‑14 — which is most of the 5.2% fall from 2012‑13 that otherwise reads as a staffing cut.
Residential care beds for the aged added for 1984‑85 to 2009‑10, from Statistics Canada’s residential care survey — beds staffed and in operation, the same measure used for hospital beds. Kept as its own series rather than as the long‑term care line extended backwards: the survey’s “homes for the aged” is a wider group than Ontario’s licensed long‑term care homes, running a steady 1.16 to 1.26 times the FAO’s rate across the fourteen years both cover, and the build now checks that wedge stays put. On its own terms it says the fall was well under way before the FAO series begins: 166 beds per 1,000 aged 75+ in 1984‑85, 140 by 1990‑91, 117 by 1996‑97.
The lines that still begin in 1990 were searched for earlier sources, and what was found — including what was ruled out — is written down on the Sources & gaps tab rather than left as a shrug. The bed‑by‑type split does exist back to 1976, in the same Statistics Canada volumes the bed totals come from, but as percentages by type of unit rather than counts in Ontario’s four categories. Historical Statistics of Canada, the obvious place to look after it gave up the 1866 school series, turns out to be a dead end for beds: national totals only, and rated capacity rather than staffed beds.
Hospital beds now begin at 1981‑82. Beds staffed and in operation from Statistics Canada’s Hospital Annual Statistics, continuing into the Ontario Health Coalition’s counts from 1990. The two overlap in 1990, 1991 and 1992 and agree to about a tenth of a per cent, which is what allows them to share a line.
Removed the rated‑bed and approved‑bed series added earlier the same day. Both are permissions — what a hospital was approved to run — rather than beds with staff beside them, and they do not belong next to the counts that are.
Physicians now begin at 1971, from CIHI’s Scott’s Medical Database. The rate recomputed from those counts tracks the FAO’s published series to a mean of 0.7 per 100,000 across the 34 years both cover.
Ontario population now begins at 1951, from the quarterly estimates.
Public‑school enrolment and full‑time teachers added for 1866‑67 to 1975‑76, from Historical Statistics of Canada. Kept as their own series: they never overlap the modern survey, so there is nothing to check a join against.
Prison capacity declared with no data behind it, so the subject appears where a reader would look for it instead of being silently absent.

Interface

The page no longer asks anyone else for anything. The two webfonts are served from this site rather than from Google, and the analytics tag is gone. That also removes the stylesheet that used to block the first paint while two DNS lookups and two TLS handshakes completed.
The data moved to the end of the document. A quarter of a megabyte of JSON used to sit between the stylesheet and the first element, so nothing could be drawn until all of it had arrived; the page now paints after 56 KB instead of 269.
A chart can now be shaded by which party was governing Ontario, hatched where that party held fewer than half the seats — Context → Governments, off until you ask for it. It covers all 44 parliaments since Confederation. The bands are context for reading the years and not evidence about them, which the chart says whenever they are on: capacity moves on budgets, demographics and decisions taken long before, and often by somebody else.
Whether a government held a majority is worked out from the seats rather than copied from a list, because the lists disagree. Ontario has had more minority governments than most summaries show — 1919, 1943, 1975, 1977, 1985 twice over as one government fell and another formed without an election, and 2011.
“Two axes” is now offered for series that share a unit but not a scale, which is the other half of letting counts of people share an axis. Ontario population beside educators is fifteen million against a quarter of a million, and on one scale the educators are a flat line along the bottom; the mode splits them when the spread from the largest series to the smallest passes ten to one, breaking at the widest gap so each side stays as tight as it can. It is offered and never taken — one axis can always draw them, so the chart stays on Actual values until you ask for the second scale, and the chart that has the problem is where the offer appears: a note saying one line is flat and how far apart the scales are, with a button that makes the switch. When both axes carry the same unit the notice says so outright, and each axis is titled with its series rather than with the unit, because “people” opposite “educators” reads as two quantities when the whole point is that they are one.
The picker’s subheadings and series are indented under their group heading, so what belongs to what is visible without reading the type sizes.
A Coverage tab: every subject in the database and the years it actually reports, one strip per subject against a common timeline, with the holes inside each run spelled out. It is counted off the observations on every render, so it says what is here; the Gaps table still says what is missing and where to get it.
Every series now says what it means in plain words under the chart, before its citation — what a staffed bed is, what an educator is, why a denominator was chosen.
The picker is grouped by denominator and subdivided by subject, with the same subheadings in the same order under the rates and the counts. Ontario population leads it and is what the page opens on.
Groups and subheadings collapse and remember their state. Everything starts closed except the section holding the series on screen.
Each small‑multiple panel carries an × to drop that series.
The year range follows the selection until you set it yourself, so picking a series that starts in 1866 no longer squashes everything else into a corner.

Foundations

The build checks itself five ways on every run: period alignment against the FAO’s published bed rate, educators against teacher full‑time equivalents, physicians against the FAO’s published rate, staffed beds against the Ontario Health Coalition, and residential care against long‑term care. It warns rather than fails, and prints all five.
That last check is there to hold a negative result in place. Residential care and long‑term care share a denominator and look like the same measurement, so the standing temptation is to splice them into one line reaching back to 1984. The reason not to is how steady the gap between them is — a definition does that, drift does not — and the build now says so on every run instead of leaving it to a comment nobody reads.
The pipeline reads committed CSVs and nothing else; the source workbook that used to sit between the references and the data was removed.
First build of the database and this explorer, published as a single static page with its data inlined.