Spatiotemporal Tracking Pipeline
I compare four filters on football tracking data. I check how they change speed estimates and Spearman pitch-control maps.
I'm Hritik, a CS student in Hyderabad. I work on football tracking data and spatial indexing.
These are two small simulations I built to explain ideas from my research. The data is made up for illustration; it is not from real matches.
Noisy tracking positions can give the wrong speeds. The Savitzky-Golay filter shows a cleaner path. Use the Raw/Filtered toggle to compare the noisy path with the filtered one. Read this project.
Move your cursor over the field.
A quadtree splits space into smaller boxes where points are dense. A range query can then check nearby boxes instead of every point. Move the cursor, or tap on a touch screen, to move the query circle.
Tap to move the circle.
I compare four filters on football tracking data. I check how they change speed estimates and Spearman pitch-control maps.
I built Synapse OS to watch what groups of AI agents are doing. It shows their connections, costs, timing, and risk scores. It can also stop a task and ask a person to approve it.
I use Python and public StatsBomb data to make pitch plots, pass maps, shot maps, xG charts, heatmaps, and spatial-control studies.
I'm building xAssassin to study space creation, off-ball movement, and pressing triggers from tracking data. It is still unfinished.