The Geography of Choice

Annex D. Cosine similarity between GPS-based and census-based commuting flows

Figure A D.1. In Australia, GPS-based and census-based commuting flows are consistent

Cosine similarity indices in Local Government Areas between Habidatum mobile-phone-based home-work flows and ABS census-based commuting flows

Source: OECD computations based on Habidatum mobile phone data and ABS census-based commuting flows

Figure A D.2. Although sparse, Canadian GPS-based commuting flows are relatively consistent with census-based flows, especially in cities

Cosine similarity indices in municipalities between Habidatum mobile-phone-based home-work flows and Statistics Canada census-based commuting flows

Source: OECD computations based on Habidatum mobile phone data and Statistics Canada census-based commuting flows

Figure A D.3. In Japan, GPS-based and census-based commuting flows are consistent

Cosine similarity indices in municipalities between Habidatum mobile-phone-based home-work flows and Official Statistics of Japan (e-Stat) census-based commuting flows

Source: OECD computations based on Habidatum mobile phone data and Official Statistics of Japan (e-Stat) census-based commuting flows

Figure A D.4. In the UK, GPS-based and census-based commuting flows are consistent

Cosine similarity indices in municipalities between Habidatum mobile-phone-based home-work flows and ONS/NISRA/Scotland’s census-based commuting flows

Source OECD computations based on Habidatum mobile phone data and census-based commuting flows collected from ONS for England and Wales, NISRA for Northern Ireland and Scotland’s census for Scotland.