tppllm/chicago-crime

Chicago Crime Dataset

This dataset contains reported crime incidents in Chicago from January 1, 2022, to December 31, 2023. It includes 4,033 sequences with 202,333 events across 20 crime types. The data is sourced from the Chicago Data Portal under the Terms of Use. The detailed data preprocessing steps used to create this dataset can be found in the TPP-LLM paper and TPP-LLM-Embedding paper.

Update (2025-10-28): Added three timestamp fields (timestamp_event, timestamp_since_start, timestamp_since_last_event) in seconds, and refreshed the dataset using the latest official source data. If you wish to reproduce the original TPP-LLM results, please refer to the earlier dataset snapshots in the Commit History.

If you find this dataset useful, we kindly invite you to cite the following papers:

@article{liu2024tppllmm,
  title={TPP-LLM: Modeling Temporal Point Processes by Efficiently Fine-Tuning Large Language Models},
  author={Liu, Zefang and Quan, Yinzhu},
  journal={arXiv preprint arXiv:2410.02062},
  year={2024}
}

@article{liu2024efficient,
  title={Efficient Retrieval of Temporal Event Sequences from Textual Descriptions},
  author={Liu, Zefang and Quan, Yinzhu},
  journal={arXiv preprint arXiv:2410.14043},
  year={2024}
}