Traffic Flow Forecasting from Open Sensor Data
Predict traffic speed and volume across a road network fifteen to sixty minutes ahead using public sensor data, treating the road network as a graph rather than each sensor as an independent series. A well defined problem with excellent open datasets and an honest baseline.
Intermediate
A small team, or one strong student willing to learn something new.1 semester, 2 students
5 to ship
3 optional extrasSuggested stack
What you should ship
- Ingestion and cleaning pipeline handling missing sensor readings and outliers
- Baseline models including historical average and a simple statistical forecaster
- A graph aware model that uses road network adjacency, compared against the baselines
- Error reported at 15, 30 and 60 minute horizons, broken down by time of day
- Map interface showing predicted against actual conditions
If you have time left
- Incorporating weather and public event data as additional inputs
- Incident detection from unexpected deviation between prediction and observation
- Estimating how far a disruption propagates through the network
The problem
Congestion prediction is genuinely useful for routing, signal timing and planning, and several cities publish loop detector data openly, which removes the usual data collection obstacle.
What you build
A cleaning and ingestion pipeline, baseline forecasters, and a model that exploits the network structure by treating adjacent road segments as connected nodes rather than independent series.
Why the graph structure matters
Traffic on one segment is caused by traffic on the segments feeding into it, a few minutes earlier. A model that knows the adjacency captures propagation that per sensor models cannot, and demonstrating that improvement over a strong baseline is the core result.
The baseline you must beat
Historical average by time of day and day of week. It is embarrassingly effective at this task, and a project reporting an impressive sounding error without comparing to it has not shown anything. Build it first, in week two.
The hard part
Missing data. Real sensors fail, and how you impute gaps materially changes your results. Document the strategy and show its effect.
Scope warning
One city, one road network, one prediction target. Do not attempt to build a routing application on top; the forecasting and its evaluation is the project.
Ideas and guidance, not finished projects
These are project ideas and scoping guidance, published free for students to use as a starting point. I do not build, write, or sell final-year projects, and I do not complete coursework for anyone. Take an idea, make it yours, and build it.
Not sure this one fits you?
I help students pick an idea that matches their skills and their deadline, then scope it down until it is finishable. Guidance only, never done for you.
Stuck on something specific?
Writing only gets you so far. If you want an answer to your situation rather than the general case, book a session and we will work through it together. Every session is free; a few slots open each week.
Follow along
New writing, resources and project ideas land here first.