How many people here?

Dit is een foto van een winkelstraat met mensen 

Statistics Netherlands (CBS) is constantly studying the usability of new data sources. One such source are mobile phone cellular network data. These data can be used to produce estimates of the so-called daytime population, providing insight into the number of people located in a particular area at a particular time.

CBS produces monthly population figures for the Netherlands in a regular publication. These are based on municipal population registers. Using the new data presented here, CBS can also produce figures on the population count at particular times in particular areas and compare these figures to the number of residents based on home address registrations. The attached animation shows an example of how these big data may be used. CBS processes around 15 billion records a month in the production of these figures.

Privacy guaranteed

These figures are based on mobile telephone cellular data. As mobile phone devices make contact with transmission masts, the location of these devices can be (globally) determined. CBS uses the data to estimate the number of persons present in a certain area. Raw data are sourced from Vodafone and delivered by Mezuro, which specialises in processing and analysis of big data derived from the Vodafone mobile phone cellular network.

CBS only receives anonymised, aggregated data on the number of mobile phone devices. Privacy is guaranteed: no personal data are supplied and individual users cannot be identified or traced. CBS is therefore unable to establish the location of any individual person at any time. The data are solely used to measure the number of people at a given location without being able to establishing these people’s identities.

Commuters and visitors

The animation shows the number of people in a given area relative to the official number of residents in that same area, and the hourly changes in that proportion. Areas turn blue when the total number of people at that time is lower than the official population. When they change to red, the number of people has exceeded the local population size. A yellow area indicates that the total number of people present is equal to the number of residents. Area sizes correspond to municipalities. The city centres of the five largest cities – Amsterdam, Rotterdam, The Hague, Utrecht and Eindhoven – are marked as separate areas.

The data in the animation relate to the period from Monday 6 May 2013 to Sunday 12 May 2013. The year 2013 was chosen to ensure the data have been thoroughly checked. The currency of information is not relevant in this prototype.

Areas turning blue during the day are generally commuter areas; many residents leave these areas to go to work elsewhere. Areas coloured red during daytime indicate a density increase mainly related to employment in these areas. In yellow-coloured areas, there is hardly any mobility. Another explanation could be that incoming and outgoing population flows are nearly equal in size. 

Clearly visible in the animation are the municipalities in the Randstad urban conglomeration with many commuters as residents. Far fewer people are shown during daytime than are registered as residents. One such municipality is Almere. Amsterdam, on the other hand, has many more people during daytime than the number of residents.

Thursday 9 May 2013 was Ascension Day, a public holiday in the Netherlands. Areas indicating larger numbers of people on that day are very different from normal weekdays. More people are seen in the coastal municipalities, while city centres and municipalities with amusement parks are more crowded as well, e.g. Loon op Zand with the Efteling theme park.

Applications

These data may be used by CBS as a source for figures on mobility, security and more. For example, policymakers may use this information to plan public facilities and draw up evacuation plans. The data could also be combined with other information, e.g. on traffic hubs, events or population composition. In all such applications, reporting of absolute numbers has more relevance than relative numbers - comparing daytime populations and residents - as in the visualisation.

Feedback welcome

We would like to know your opinion on this new daytime population statistic, whether it is just your general impression, a specific issue or a positive element. For example, do you think the information is presented clearly? If not, how could it be improved? Is any relevant information missing?