MobilityMirror

数据集MobilityMirror由华盛顿大学信息学院和德雷塞尔大学计算与信息学院联合创建，旨在通过发布移动数据支持综合交通研究。该数据集包含197,049个独立骑行者的详细信息，包括骑行起始和结束位置，以及性别和是否使用头盔等。创建过程中，研究人员采用了基于因果关系的方法来移除数据中的特定偏见，确保数据在不同应用场景下的公平性和准确性。MobilityMirror数据集的应用领域广泛，包括模型训练、需求评估、收入估计和交通趋势分析等，旨在解决交通数据中的偏见问题，提高模型的公平性和准确性。

The MobilityMirror dataset was jointly created by the School of Information at the University of Washington and the College of Computing & Informatics at Drexel University, aiming to support comprehensive transportation research by publishing mobile travel data. This dataset includes detailed information of 197,049 individual riders, covering trip start and end locations, gender, and helmet usage status, among other attributes. During the dataset development process, researchers adopted a causality-based approach to remove specific biases in the data, ensuring its fairness and accuracy across diverse application scenarios. The MobilityMirror dataset has a wide range of applications, including model training, demand assessment, revenue estimation, and transportation trend analysis, among others, and it is designed to address bias issues in transportation data and improve the fairness and accuracy of relevant models.