This project delves into the capabilities of MT-GNN, a state-of-the-art model for harnessing the power of graph learning in spatio-temporal analysis. To gauge its effectiveness, we devised several baseline models and pitted them against MT-GNN in a head-to-head comparison.
Understanding the intricacies of MT-GNN, with its multifaceted components and complex interactions, was pivotal to our investigation. We meticulously studied the model and its accompanying code, gaining a deep appreciation for its inner workings.
Our journey took us across diverse datasets, encompassing traffic patterns, solar energy production, air quality measurements, urban mobility trends, and energy consumption behaviors. We meticulously analyzed the results, unveiling MT-GNN's strengths and potential areas for further exploration.
This project offers valuable insights for those seeking to leverage the power of MT-GNN for their own spatio-temporal modeling endeavors.