I Drove Uber for 3 Years. Then I Analyzed It.

3,448 trips. $70,768 earned. $31,323 taken by Uber. One question: where was the real money?

Live Demo → HuggingFace Spaces

---

The Story

From May 2022 to May 2025, I drove Uber full-time across Los Angeles while building my data engineering skills on the side. Every trip, every surge, every 5-star rating — all logged by Uber's system.

At some point I realized: I was sitting on 3 years of personal business data and had never actually analyzed it. So I requested my full export, built an ingestion pipeline, modeled it in PostgreSQL, and turned it into a dashboard. What I found surprised me.

---

What I Found

Short trips pay more than long ones — 0–2 mile trips average $118/hr. Trips over 20 miles drop to $80/hr. Uber's minimum fare kicks in on shorts, and you fit more of them per hour. Counterintuitive until you do the math.

Halloween > New Year's Eve — October 28, 2023 was my single best day: $418, 10 surge trips. January 1st had the highest avg fare per trip ($45) but fewer trips overall. Event volume beats event pricing.

Uber took $14,303 in commission — then paid back $6,362 in incentives — net commission burden after incentives: ~18%, not the 32.3% gross rate. The incentive programs partially offset what looks like an aggressive take rate.

I earned 54% more per hour in 2025 than 2022 — $68/hr → $105/hr, while avg trip distance dropped. Shorter trips, smarter routing, better surge timing.

98.9% of my ratings were 5-star — 1,669 rated trips, 1,651 at 5 stars. Not a vanity metric — high ratings unlock better surge zones and priority dispatch in LA.

---

The Numbers

Metric	Value
Total gross earned	$70,768
Net after commission	$39,445
Uber's gross take rate	32.3%
Net take rate (after incentives)	~18%
Avg earnings per hour	$87/hr gross
Best trip type	Short 0–2 mi → $118/hr
5-star rating rate	98.9% (1,669 ratings)
Surge trips	231 trips · $1,531 earned
Best single day	Oct 28 2023 → $418
$/hr growth 2022–2025	$68 → $105 (+54%)

---

Screenshots

📊 Overview — KPI Cards & Year-over-Year

💰 Earnings — Monthly Timeline & Commission Breakdown

🗺️ Trips — Heatmap, Distance & Surge Analysis

⭐ Ratings & Tips — Distribution & Tips by Year

---

How It Works

Uber CSV export (3 files · personal data)
        ↓  Python ingestion
   Supabase (PostgreSQL cloud)
        ↓  SQL query layer (12 analytical scripts)
   Dashboard (4 pages · Plotly)
        ↓  Docker
   HuggingFace Spaces (always-on)

Layer	Tool
Data source	Uber driver CSV export (personal)
Database	Supabase (PostgreSQL cloud)
Ingestion	Python + pandas + REST API
Analytics	12 SQL scripts — window functions, CTEs, cohort analysis
Dashboard	Streamlit + Plotly
Containerization	Docker + Docker Compose
Deployment	HuggingFace Spaces (Docker SDK)

---

Dashboard Pages

Overview — 8 KPI cards · year-over-year comparison · "Where the Money Goes" stacked bar

Earnings — Monthly gross/net timeline · trips per month · avg fare trend · commission breakdown by year

Trips — Hour × weekday heatmap · distance bucket analysis ($/hr by range) · surge vs regular · airport vs regular

Ratings & Tips — 5-star distribution · tips by year

---

Quick Start

Option A — Docker

git clone https://github.com/evgeniimatveev/uber-driver-analytics.git
cd uber-driver-analytics
cp .env.example .env        # fill in your Supabase credentials
docker compose up

Open http://localhost:8501

Option B — Python

git clone https://github.com/evgeniimatveev/uber-driver-analytics.git
cd uber-driver-analytics
pip install -r requirements.txt
cp .env.example .env
python -m streamlit run dashboard/app.py

---

Run It With Your Own Data

Uber drivers can export their full trip history at Uber Help → Request Your Data.

1. Place the 3 CSVs in data/:
   • driver_lifetime_trips.csv
   • driver_payments.csv
   • driver_lifetime_ratings_received.csv
2. Create a free project at supabase.com
3. Run sql/schema.sql in Supabase SQL Editor
4. Load data: python ingestion/load_supabase.py
5. Launch: python -m streamlit run dashboard/app.py

---

Project Structure

uber-driver-analytics/
├── data/                    # CSVs here (gitignored — personal data)
├── sql/
│   ├── schema.sql           # PostgreSQL table definitions
│   └── analysis/            # 12 analytical SQL scripts
├── ingestion/
│   ├── load_data.py         # CSV → local PostgreSQL
│   └── load_supabase.py     # CSV → Supabase via REST API
├── dashboard/
│   ├── app.py               # Streamlit app (4 pages)
│   └── db.py                # SQL query layer
├── Dockerfile
├── docker-compose.yml
├── requirements.txt
└── .env.example

---

Data Schema

Table	Rows	Description
trips	3,745	Every trip — timestamps, fares, distance, surge flag
payments	21,112	Per-trip payment breakdown by category
ratings	1,669	5-star ratings received

---

Availability

Layer	Detail
Hosting	HuggingFace Spaces (Docker SDK)
Uptime	24/7 — HF Spaces does not sleep
Database	Supabase (PostgreSQL) via HF Space secrets
Keepalive	GitHub Actions every 30 min · cron-job.org backup

---

Personal Uber driver data · Los Angeles, CA · May 2022 – May 2025 · Built by Evgenii Matveev