Synthetic Datasets for Pedestrian and Vehicle Movement Analysis During the Arbaeen Pilgrimage Using Python-Based Simulation

Overview This article introduces two synthetic datasets generated for the purpose of analyzing crowd and traffic behaviors during the Arbaeen pilgrimage. The data was programmatically created using Python to simulate realistic pedestrian and vehicle movement scenarios in key Iraqi provinces during the pilgrimage season. These datasets serve as the foundation for the research paper titled “Secure System for Anomaly Detection and Data Analysis in the Arbaeen Pilgrimage Using ML Algorithms”, which was submitted to the Arbaeen Conference 2025. Motivation The Arbaeen pilgrimage attracts millions of visitors, creating complex traffic and pedestrian dynamics that require effective monitoring and anomaly detection for safety, logistics, and infrastructure planning. However, obtaining large-scale, annotated real-world data under such conditions is challenging. To address this, we developed simulated datasets to mimic realistic behaviors observed during this religious event. Dataset Description 1. Pedestrian Dataset (pedestrian_data_en_3000.xlsx) This dataset contains 3,000 records representing pedestrian activity across five provinces: Karbala, Najaf, Babil, Diwaniya, and Wasit, in areas such as city entrances, squares, and streets near major shrines. Key Attributes Include: (Date & Time of observation, Expected vs Actual Visitor Count, Average Walking Speed and its evaluation (normal/abnormal), Difference Note indicating if the variance in counts is significant, Field Behavior (e.g., crowding, reverse flow, gatherings), Movement Type (normal or suspicious), Stop Ratio and Additional Notes (e.g., prolonged stop, surveillance)) The synthetic logic ensures realistic fluctuations in crowd sizes, walking speeds, and unusual behaviors. 2. Vehicle Dataset (vehicle_data_en_3000.xlsx) Comprising another 3,000 records, this dataset simulates vehicular activity in the same geographic scope, covering roads and garages frequently used during the Arbaeen event. Key Attributes Include: (Date & Time, Province, and Area, Vehicle Type (bus, private car, truck, motorcycle), Expected vs Actual Vehicle Count, Average Speed, Speed Note, and Stop Duration, Field Notes (e.g., sudden stops, side-road usage), Movement Type based on combined factors, Additional Notes describing behavioral anomalies) Both datasets are structured in Microsoft Excel format and are suitable for machine learning tasks such as classification, anomaly detection, and traffic pattern prediction. Data Generation Methodology The datasets were generated using Python 3, employing the pandas, datetime, and random libraries. The scripts simulate 10 days of movement data with realistic variability, ensuring that edge cases such as high deviations, low/high speeds, and abnormal stopping patterns are present for testing ML algorithms.