Toronto ridehail prices through the week, by ward

Understanding the maps

The maps show how fares, earnings, and commission vary by time and place across the city of Toronto, in ways that a simple rate card would not. They paint an incomplete picture, however. The rider map comes from the City's data for all companies, not just Uber, while the driver and Uber commission maps come from a small number of driver account statements for the month of July 2026. Also, location (at the ward level) and hour of the week are parts of the "algorithmic pricing" regime that Uber and other platforms now use, but they are not the whole story. Uber also makes more fine-grained adjustments based on individual trip characteristics, but we do not yet have enough data to reverse engineer these. If we can collect more trip records then we can make the map more detailed and more accurate.

The maps colour each ward in the city, for each hour of the week, to show how a representative trip would be priced if it were picked up at that time and place, on average. The sample trip is 6 km long and takes 13 minutes, but the pattern is very similar for other trip distances and durations.

Monday to Wednesday are grouped together because their patterns are very similar. Thursday to Sunday are each distinct, and are shown on their own.

The numbers are calculated as follows:

• The rider's price is calculated from the City of Toronto open data, with HST removed.
• The driver earnings are calculated from data in the driver account statements, including Uber's adjustments. We have found that Uber makes some adjustments based on location, on the time of day, and on the hour of the week. It also adjusts earnings based on how far a driver has to go to pick up a passenger.

The trip is priced on the UberX rate card of the zone it starts in, because the cards differ by zone, and the panel beside the map breaks each price into three levels:

• 1 The zone's rate card and commission rate, fixed for months at a time (see Rate cards), plus the booking fee.
• 2 Pricing by time and place: Uber's fixed weekly schedule of percentage adjustments by time block and zone; Uber's average adjustments to drivers' offers and to riders' prices in that time block and zone; other pricing that follows the time and place, on average; and, for drivers, the pickup wait. Uber's commission applies to the price after the schedule, so a cut or a rise is shared between the driver and Uber in proportion to the commission rate. The adjustment to the driver's offer moves money between the driver and Uber without changing what the rider pays. It varies far more from trip to trip than the schedule does, so the panel shows only its average for the time block and zone, and its variation is part of level 3. On average it evens out the schedule, taking more where the schedule cuts little, so the two together vary less through the week than the schedule alone. The adjustment to the rider's price moves money between the rider and Uber without touching driver earnings; the rider and Uber panels show its average, and because each one is large and can go up or down, most of its variation is part of level 3. "Other" is the average of what remains beyond these lines (surge, boosts, Uber's smaller adjustments to riders' prices and, for riders, differences between the City's data and the drivers' trips). The statements do not say why, so we cannot account for it.
• 3 Trip-by-trip pricing that levels 1 and 2 don't explain, shown as the range within which most trips fall. Offers at the same time and place vary this much for reasons we cannot see.

The whole week at once

The chart below is a heat map that gives a different perspective on the same data. It shows every ward (rows, central Toronto first) and every hour of the week (columns), for the reference trip.

Other trip lengths

The map shows one trip length because the pattern across wards and hours is almost the same for any length: every value is the zone's rate card for the trip times a ratio that does not depend on the trip, plus the booking fee. The table gives the dollars over the typical week (trip-weighted, Toronto wards) for four lengths.

The booking fee does not scale with the rest of the price. It depends on the trip's distance and its pickup zone, and it rises steeply beyond 15 km, so Uber's share of the price changes little up to medium lengths and rises on long trips.

Uber's share here is of the price implied by the statement models (driver earnings, Uber's take and the HST on Uber's fees), so that both parts of it come from the drivers' trips.

On the map's trip Uber keeps $0.18 less from a central pickup than from one outside the centre; $0.06 of that is the lower central booking fee, and most of the rest is the cheaper central rate card. We are less sure of the gap on longer trips. The models give, outside less central, 3 km: +$0.25, 6 km: +$0.18, 11 km: −$0.19, 26 km: +$0.00, but they assume Uber's share of the rate card, before the booking fee, is the same at every length. The statements have only 85 central trips over 8 km, too few to confirm that assumption or the swing in the gap.

Rate cards

UberX rate cards from 2026-07-04 to 2026-07-31, the most common for each pickup zone in the drivers' statements; the commission is the median effective rate on that card. The booking fee is set by the trip's distance and pickup zone, not its duration, its fare or the hour: a flat fee up to 5 km, then a rate per km, then a steeper rate beyond 15 km, up to a cap of $10.15. These are our estimates from the drivers' statements (see algorithmic_notes.qmd).

Where the numbers come from

• Riders: The map shows the typical fare that riders pay for trips at the listed time and starting from each ward. The fares are calculated based on the City of Toronto's open data, from 2026-07-04 to 2026-07-31, with HST, tips, and fixed fees removed. The City data includes 6,579,816 ridehail trips, including Uber, Lyft and Hopp trips. The City reports average prices between wards for each hour, and the value is scaled to match the reference trip. median across the weeks, so that event nights do not dominate; they are listed under Nights that were different). Each ward-hour's price, less the booking fee, as a ratio to its zone's rate card is applied to the reference trip, and the booking fee for that trip added back. Lyft and Uber's other products dilute Uber's own pattern, and the level is within about 6% of Uber's own (see algorithmic_notes.qmd). The panel splits the price into Uber's fixed weekly schedule (known exactly), the average adjustment to the rider's price on the drivers' Uber trips (an estimate here, since the City's data also covers other companies and products), and everything else riders paid in that hour and place.
• Drivers (modelled): predicted driver earnings, as a share of the rate card, from 1,987 trips in the drivers' own statements since 2026-07-01, on the rate cards in force in the City's data: the weekly schedule (time block by central or outside pickup) plus the pickup wait, applied to each ward and hour with the open data's average wait. It explains 20% of the trip-to-trip variation: most of what drivers are paid varies trip by trip in ways no ward map can show, and the panel's range shows how much.
• Uber (modelled): what Uber keeps of the rider's price, after the driver's earnings and the HST on Uber's fees, modelled the same way without the booking fee, which is then added for the trip. It explains only 7% of the variation from trip to trip, because Uber's trip-by-trip adjustments swing single riders' prices by about a fifth either way. The rider layer is observed for all companies, while the driver and Uber layers are modelled from four drivers' Uber trips, so the three need not add up exactly; the panel gives the difference.

Limitations

Shares of the rate card are applied to trips of any length, which is close for the percentage adjustments but not exact: the pickup-wait pay is per minute, so it is a smaller share of a long trip, and minimum fares matter on short ones. Central Toronto (marked * in the heat maps) is wards 4, 9, 10, 11, 12, 13, 14 and 19, standing in for Uber's lower-commission pickup zone, which follows pickup location rather than ward lines. Pearson Airport is shown for riders only, as an approximate outline just west of the city: a trip from Pearson is priced on the airport card with its pickup fees, and the drivers' statements have too few airport pickups to model drivers' or Uber's shares there. Airport riders pay well above the airport card in the open data, beyond the colour scale (the hover text gives the value). Ward-hours with fewer than 14 trips a week in the open data (a few night hours in outer wards) are blank on the rider maps.

Driver earnings and Uber's take vary by place only between central Toronto and outside it, and with each ward's pickup wait. The rider maps show that prices also vary by ward within a zone, and offers may too, but the drivers' statements have too few trips with a known pickup location (about three per ward and time block) to show it.

The panel's difference between sources compares two measures of the rider's price: the City's open data, and what the drivers' statements imply (driver earnings, plus Uber's take, plus the tax lines deducted from drivers) for the same zone, time block and pickup wait. On the reference trip it averages −$0.07 (−1%) for central pickups and −$1.23 (−11%) outside the centre. About three quarters of its variation is a fixed offset for each ward (from about −$2 in Scarborough to about +$1 downtown), which is the ward pricing the statement models cannot see. The lower overall level outside the centre is a real disagreement between the sources, with several possible causes we cannot separate: the open data includes other companies and products; the statements cover only the trips our drivers accepted; and the conversion assumes the booking fee and airport fees are in the City's fare.

Nights that were different

The maps show a typical week: for each ward and hour, the median across the 4.0 weeks. These are the single ward-hours (with at least 150 trips) where riders paid most above that typical level, which mostly means surge pricing around events.

Model details

On the drivers' statements, Uber's weekly schedule is the line transport.fare.csp_lever, the adjustment to the driver's offer is transport.fare.dop, and the adjustment to the rider's price is transport.fare.rsp_lever.

Weighted least squares of driver earnings / rate-card fare and of Uber's take less the booking fee / rate-card fare (weights: rate-card fare). One level per time block and area, plus a step for each pickup-wait band above 3 minutes. The Uber model explains 7% of the trip-to-trip variation. The trip-by-trip ranges are the 10th and 90th percentiles of each model's residuals in the time block and area (riders: the same model of the drivers' riders' prices, less the booking fee).

Cells with fewer than 20 of the drivers' trips use their area's average: Weekdays 00:00-07:00, outside, Weekdays 00:00-07:00, central.