The repository provides the code for the data aggregation process and analysis described in High spatial resolution dataset of La Mobilière insurance customers by A. Battiston, E. Massaro, C. R. Binder, and R. Schifanella. The code is provided in the form of a Python script.
The aggregated data are available at: https://doi.org/10.6084/m9.figshare.c.5515857
- aggregation provides the code for aggregation of individual level data to municipality and zip-code level.
- analysis provides the code for the data presentation and validation
- final_checks provides the code for the final data checks performed before the submission of the dataset.
We present the La Mobilière insurance customers dataset: a 12-year-long longitudinal collection of data on policies ofcustomers of the Swiss insurance company La Mobilière. To preserve the privacy of La Mobilière customers, we propose thedata aggregated at two geographical levels, based on the place of residence of the customer: postal areas and municipalities. For each geographical area, the data provides summary statistics on: i) the demographic characteristics of the customer base,ii) characteristics of vehicles insurance policies and iii) characteristics of housing and building insurance policies. To assess thevalidity of the data, we investigate the temporal consistency of the data and the representativeness of La Mobilière customerbase along several dimensions (total population, percentage of foreigners, etc.). We also show how the insurance data canreliably model the spatial patterns of socio-economic indicators at a high-geographical resolution. We believe that the reuse ofthis data provides an opportunity for researchers to broaden the socio-economic characterization of Swiss areas beyond theuse of official data sources.