An end-to-end statistical process with mobile network data for official statistics

Abstract Mobile network data has been proven to provide a rich source of information in multiple statistical domains such as demography, tourism, urban planning, etc. However, the incorporation this routinely production official statistics is taking many efforts since diversity highly entangled issues (access, methodology, IT tools, quality, skills) must be solved beforehand. To do this, one-off studies with concrete sets are not enough and standard process put place. We propose modular structured into evolvable modules detaching strongly technological layer underlying from necessary analysis producing outputs interest. This architecture follows principles so-called ESS Reference Methodological Framework for Network Data. Each these deals different aspect source. apply hidden Markov models geolocation mobile devices, use Bayesian approach on model disambiguate devices belonging same individual, compute aggregate numbers individuals detected by telecommunication using probability theory, hierarchically integration auxiliary telco market produce final estimates number across territorial regions target population. A first simple illustrative proposal applied synthetic providing preliminary software tools accuracy indicators monitoring performance process. Currently, exercise estimation present population origin-destination matrices. an example execution comparing results simulated ground truth, thus assessing each module.