Closed DriesDeprest closed 3 months ago
This looks good to me. I like that you added the end_timestamp
.
I guess you could also consider deriving an outcome (I am thinking of "ball retention" and "forced loss") and whether or not it is a counter-pressing event. But these features are not essential and could be added later too.
I want to add support for pressure events in our data model, by adding a
PressureEvent
classIn a first phase, this event would only be created when deserializing StatsBomb data, for which I will add support. But in a later, I envision that for a given game we could generate PressureEvents out of the TrackingDataset and feed it back to the EventDataset.
Thoughts?