Synthetic Population for Île-de-France

Synthetic populations are essential for agent-based simulations, enabling applications in transport modelling, land-use planning, energy consumption analysis, and beyond. A key challenge lies in learning the underlying distribution from limited dataset to generate a larger, representative datasets. While numerous methods have been proposed, we note the lack of open-source, large-scale datasets. To address this gap, we propose this open-source and large-scale synthetic population dataset, at both the individual and household levels. These datasets are derived from the French census data, amde by the French Institute of Statistics (INSEE) and publicly available.

This dataset supports reproducible research, facilitates benchmarking, and is helpful for model selection in with limited data scenarios.

Data processing is available at this Git reposetory: https://github.com/vdarsel/French_Census_Data_Processing

合成人口（synthetic populations）是基于智能体的模拟（agent-based simulations）的核心支撑，可支撑诸如交通建模、土地利用规划、能耗分析等多元应用场景。当前面临的核心挑战在于，如何从有限数据集中学得其潜在分布，以生成规模更大且具备代表性的合成人口数据集。尽管已有诸多方法被提出，但目前仍缺乏开源的大规模合成人口数据集。为填补这一空白，我们构建了这份覆盖个体与家庭两个维度的开源大规模合成人口数据集。该数据集源自法国统计研究所（French Institute of Statistics, INSEE）公开的法国人口普查数据，且已对外公开可用。

本数据集可支撑可复现研究，便于开展基准测试，并有助于在数据有限的场景下进行模型选型。

数据处理代码可在此Git仓库获取：https://github.com/vdarsel/French_Census_Data_Processing