Datasets
This page collects public datasets that I have generated or curated through my research and teaching activities.
They are provided as resources for students, collaborators, and researchers interested in exploring areas of my work, including optimisation and integer programming.
It should be noted that each dataset entry includes a short description, download link, and (where relevant) a citation to associated publications. If you make use of any of these datasets, please acknowledge the source appropriately.
The accompanying images provide a visual introduction to each dataset. The maps show the exact locations recorded in each dataset.
Datasets
-
Bin Packing Benchmark Dataset
A. Williams (2026)
[Zenodo]
-
Set Covering Benchmark Dataset
A. Williams (2026)
[Zenodo]
-
Sudoku Clue-Count Benchmark Dataset
A. Williams (2026)
[Zenodo]
-
Classical Convex Bodies Dataset
-
0/1 Knapsack Benchmark Dataset
A. Williams (2026)
[Zenodo]
-
Frobenius Number Benchmark Dataset for Primitive Integer Vectors
A. Williams (2026)
[Zenodo]
-
Explore Wales: A Geospatial Dataset of Heritage and Visitor Locations
-
Formula 1 World Championship Calendar and Circuit Data (2026)
-
London West End Theatres: Geospatial, Historical and Venue Data
-
MAX-SAT Incidence Forest Benchmark Dataset
A. Williams (2026)
[Zenodo]
-
Formula 1 Qualifying and Race Results: Ground Effect Era (2022–2025)
A. Williams (2026)
[Zenodo]
-
Formula 1 Qualifying and Race Results: Hybrid Era (2014–2021)
A. Williams (2026)
[Zenodo]
-
London Underground Stations: Geospatial, Passenger Usage and Accessibility Data
-
London Landmarks: A Geospatial Dataset for Urban Exploration and Route Optimisation
-
Royal Parks Points of Interest: A Geospatial Dataset for Mapping and Route Optimisation
-
Ted Lasso Filming Locations: A Geospatial Dataset for Route Optimisation and Analysis
-
Totally Unimodular Node-Arc Incidence Matrices: Large Collection (50,000 - 100,000 nodes)
A. Williams (2025)
[Zenodo]
-
Totally Unimodular Node-Arc Incidence Matrices: Medium Collection B (20,000 - 50,000 nodes)
A. Williams (2025)
[Zenodo]
-
Totally Unimodular Node-Arc Incidence Matrices: Medium Collection A (1000 - 20,000 nodes)
A. Williams (2025)
[Zenodo]
-
Totally Unimodular Node-Arc Incidence Matrices: Small Collection (50 - 1000 nodes)
A. Williams (2025)
[Zenodo]