“The problem with existing big data models however is that they do not capture what geographical structures such as neighbourhoods, towns, cities, regions, countries etc. This makes it difficult, for example, to generate good models for future mobility. Intuitively it seems wrong that you cannot see, for example, that a neighbourhood or urban zone has a typical area. The model uses anonymized mobile data from more than 700,000 individuals worldwide, and identifies scales—neighbourhoods, towns, cities, regions, countries—for each person based on their movement data. This has implications for some of society’s most important decisions:“Better models of mobility are important.
