When Context Compensates for Sparse Event History: AlphaEarth for Spatio-Temporal Point-Process Forecasting
For practitioners of spatio-temporal forecasting with limited local data, this shows that spatial context can substantially improve transfer performance.
The paper investigates whether exogenous spatial context (AlphaEarth embeddings) can compensate for sparse event histories in spatio-temporal point-process forecasting. On EMS data across eight held-out regions, context improves out-of-region predictions by 2–6× at 1–2 weeks of history, tapering to 10–20% at 20–104 weeks.
Spatio-temporal point-process models must often generalise across space when local event histories are sparse. We study whether exogenous spatial context can compensate in such regimes. Using a fixed log-Gaussian Cox process backbone, we compare an event-only model with the same model augmented by AlphaEarth embeddings as linear spatial context. We evaluate spatial transfer on emergency medical services (EMS) forecasting across eight held-out regions, fixed forecast anchors, and a sweep over history length $w$, using only AlphaEarth (AE) embeddings available strictly before each anchor. AE improves out-of-region predictive performance across all history regimes, with the largest gains under scarce histories: approximately $2$--$6\times$ multiplicative improvements at $1-2$ weeks, tapering to roughly $10$--$20\%$ at $w=20$--$104$ weeks. These results show that contextual information can substantially stabilise spatially transferred point-process forecasts when event history is limited.