An activity-based synthetic population of Gothenburg, Sweden: Dataset of residents in neighbourhoods

Abstract

A synthetic population is a distribution of synthetic agents that replicates the demographic distribution of a real-world population based on census records. This paper presents an end-to-end model to generate a synthetic population of residents in Gothenburg, Sweden, along with activity schedules and mobility patterns for present and past populations. Using a stochastic modelling approach, we describe the model and present its corresponding dataset. The model is designed for applications in neighbourhood planning and includes detailed replicas of people in different neighbourhoods of Gothenburg organised as persons, households, houses, buildings, and daily activity chains. While the persons, households, and houses are synthetic replicas, they are connected to existing buildings. The model considers the allocation of primary and secondary locations based on a gravity model, realistic routing for active, public, and private motorised modes of transportation and allows users to introduce new buildings and amenities if needed. The model aims to impute national-level mobility patterns from a household travel survey and apply them locally to capture the nuances of a neighbourhood's built environment and demographic composition.