What rideshare really pays
I reconciled three and a half years of my own driving data to answer one question: per hour of my time, what does this job actually pay? The app makes it look like about $95. The real number is $41.61.
- Trips
- 25,549
- Hours on the clock
- 7,876
- Miles
- 158,940
- Net earnings
- $327,729
- Days driven
- 1,040
- Raw data
- 415,000+ rows
The app's number leaves out half the job
Uber shows drivers what they earned while a rider was in the car. That skips the drive to the pickup, the wait in the airport lot, the empty drive home from a long trip, and the dead time between requests. All of that is still work, and none of it is paid.
I ran Gridwise the whole time, which logs every hour I was on the clock. Divide what I actually netted by those hours and the picture changes.
The app's number is about 2.3 times too high. That's not a lie, it's measuring the wrong thing. If you only ever look at the app, you'll badly overestimate what an hour of this work is worth.
Getting four sources to agree
No single export told the whole story, and they didn't agree with each other. Most of the work was making them line up.
- Pull everythingUber's trip history (46,548 rows) and payment ledger (368,869 line items), Lyft's rides, sessions and payouts, and Gridwise's daily earnings, hours and mileage. Six files, 415,000+ rows.
- Reconcile the ledgersGridwise's earnings totals matched Uber's payment ledger to within $1,000 over three years. The biggest single gap came down to one day.
- Explain the outliersMy best day on paper was February 28, 2024: $1,920.98 on four rides. It turned out to be a $1,868.16 EV settlement Uber paid outside its normal earnings records. It's real money, so it stays in the totals, and it's flagged on every chart.
- Use the right unitEverything is per hour on the clock, per mile and per trip, net of platform fees. Not per hour with a rider in the car.
| Year | Days | Trips | Net earnings | Hours | Miles | Per hour | IRS mileage deduction |
|---|---|---|---|---|---|---|---|
| 2023 | 306 | 7,565 | $94,277 | 2,242.6 | 44,730 | $42.04 | $29,299 |
| 2024 | 312 | 7,595 | $97,906 | 2,283.9 | 46,882 | $42.87 | $31,411 |
| 2025 | 312 | 7,718 | $98,753 | 2,462.5 | 49,754 | $40.10 | $34,828 |
| 2026 so far | 110 | 2,671 | $36,794 | 887.4 | 17,573 | $41.46 | $12,741 |
| Total | 1,040 | 25,549 | $327,729 | 7,876.4 | 158,940 | $41.61 | $108,278 |
The deduction column is miles driven times the IRS standard mileage rate for that year. Other per-unit numbers for the whole period: $2.06 per mile, $12.83 per trip, 3.24 rides per hour, 20.2 mph average.
Steady, month after month
Outside the settlement month, every full month landed between about $6,500 and $9,600. The hourly rate barely moved year to year: $42.04, $42.87, $40.10, and $41.46 so far in 2026. The job is consistent. The question is how to make it pay more.
When I drive matters more than how much
Saturday paid $53.24 an hour. Monday paid $33.11. Same car, same driver, same city, and 61% more for every hour. Per day the gap is just as wide: $453 on an average Saturday against $272 on a Tuesday. It's the biggest lever in the whole dataset.
The hours tell the same story. 97% of my Uber trips started between 3 PM and midnight, heaviest on Friday and Saturday nights. Mornings are basically empty. That's my schedule, not the market's, so it says as much about how I work as where the demand is.
Uber's cut went up in 2026
From Uber's per-line-item payment ledger, I divided Uber's total commission by everything riders paid. For three years it held right around 40%. In 2026 so far it's 44%. My hourly rate held steady through it, which means I'm working smarter to stay even, not getting ahead.
Surge is a bonus, not a plan
71% of my Uber trips had no surge at all, and those base-price trips are the backbone of the income. Surge pays well when it shows up. The 248 trips at 3x or higher averaged $32.19, about 2.2 times a normal $14.67 fare. But it's rare, and chasing it means driving empty to get there.
Three smaller findings
Airport runs pay, and I skip them
Pittsburgh International trips averaged $42.69 against $15.76 for a city trip. Only 139 of my trips were airport runs, 0.6%. Doubling that would add roughly $1,100 a year, before counting the longer drive and the empty trip back.
Riders tip when the ride is good
Riders tipped on 39.1% of my trips, $35,911 in total. That's money that only shows up when the ride goes right.
Nearly all Uber
97% of my trips were Uber. Lyft paid almost the same per trip, $12.35 against $12.77, across 768 rides. It's a real backup option I've barely used.
Why I did this
Most portfolio projects start with a clean dataset somebody else prepared. This one started with six files that disagreed, an outlier day that looked like an error, and a headline number that was technically true and still wrong. Sorting that out is most of what analytics work is in a front office or anywhere else: find the right source, reconcile what doesn't match, pick the unit that answers the real question, and explain it so someone can act on it.
About the data
- Sources: Uber driver_lifetime_trips (46,548 rows), Uber driver_payments (368,869 line items), Lyft rides, sessions and payments, and Gridwise daily earnings and mileage.
- Period: January 1, 2023 to May 9, 2026. 1,040 days driven, 7,876 hours, 158,940 miles.
- All earnings are net of platform fees. Headline figures are totals from the exports, not estimates.
- Built with Python and pandas. The summary numbers behind every chart are in data/summary.json, and build.py turns them into this page. The raw exports stay private because they're my personal financial records.
Want to talk about it?
I'm looking for a front office or analytics role in the Pittsburgh area. I spent 15 years running minor league sports businesses before I went back for my master's in data analytics.