Evaluating Spatio-Temporal Forecasting Trade-offs Between Graph Neural Networks and Foundation Models

TL;DR

This study systematically compares classical, STGNN, and Time-Series Foundation Models for short-term spatiotemporal temperature forecasting in a 25-node IoT network, varying sampling frequency and sensor density. It finds that multivariate TSFMs like Moirai best capture cross-sensor dependencies, while STGNNs excel when spatial correlations are strong or data are sparse, and univariate TSFMs struggle without true multivariate integration. Enhancements such as spatial ensemble blending improve TSFMs like TimesFM under certain conditions, but architecture matters most. The work provides actionable guidance for choosing forecasting models based on data topology and sensor deployment, and releases open-source code for reproducibility.

Abstract

Modern IoT deployments for environmental sensing produce high volume spatiotemporal data to support downstream tasks such as forecasting, typically powered by machine learning models. While existing filtering and strategic deployment techniques optimize collected data volume at the edge, they overlook how variations in sampling frequencies and spatial coverage affect downstream model performance. In many forecasting models, incorporating data from additional sensors denoise predictions by providing broader spatial contexts. This interplay between sampling frequency, spatial coverage and different forecasting model architectures remain underexplored. This work presents a systematic study of forecasting models - classical models (VAR), neural networks (GRU, Transformer), spatio-temporal graph neural networks (STGNNs), and time series foundation models (TSFMs: Chronos Moirai, TimesFM) under varying spatial sensor nodes density and sampling intervals using real-world temperature data in a wireless sensor network. Our results show that STGNNs are effective when sensor deployments are sparse and sampling rate is moderate, leveraging spatial correlations via encoded graph structure to compensate for limited coverage. In contrast, TSFMs perform competitively at high frequencies but degrade when spatial coverage from neighboring sensors is reduced. Crucially, the multivariate TSFM Moirai outperforms all models by natively learning cross-sensor dependencies. These findings offer actionable insights for building efficient forecasting pipelines in spatio-temporal systems. All code for model configurations, training, dataset, and logs are open-sourced for reproducibility: https://github.com/UIUC-MONET-Projects/Benchmarking-Spatiotemporal-Forecast-Models

Figures (6)

• Figure 4.1: Adjacency matrix heatmaps for $p=20\%$ (left) and $p=60\%$ (right) in STGNNs.
• Figure A.1: Annotated tower locations (lat--lon).
• Figure A.2: Hourly temperature overlap for colocated Towers 34 and 35.
• Figure B.1: (a,b) Spatial clustering and nearest-centroid selection.
• Figure B.2: (c,d) Iterative merging of clusters by centroid distance.