MoTime: A Dataset Suite for Multimodal Time Series Forecasting

TL;DR

MoTime introduces a comprehensive suite of multimodal time series datasets across various domains, enabling evaluation of external modalities' impact on forecasting performance in both standard and cold-start scenarios.

Contribution

The paper presents MoTime, a new dataset suite for multimodal time series forecasting, facilitating structured evaluation of modality utility in diverse real-world scenarios.

Findings

External modalities improve forecasting accuracy in multiple datasets.

Short series benefit significantly from external modalities.

Impact of modalities varies with data characteristics.

Abstract

While multimodal data sources are increasingly available from real-world forecasting, most existing research remains on unimodal time series. In this work, we present MoTime, a suite of multimodal time series forecasting datasets that pair temporal signals with external modalities such as text, metadata, and images. Covering diverse domains, MoTime supports structured evaluation of modality utility under two scenarios: 1) the common forecasting task, where varying-length history is available, and 2) cold-start forecasting, where no historical data is available. Experiments show that external modalities can improve forecasting performance in both scenarios, with particularly strong benefits for short series in some datasets, though the impact varies depending on data characteristics. By making datasets and findings publicly available, we aim to support more comprehensive and realistic benchmarks in future multimodal time series forecasting research.