Anonymised aggregated mobile phone meta data may serve as a basis for indirect estimates of the number of residents that are present in their municipality and the number of visitors to this municipality at a particular time. This is the conclusion that can be drawn from the pilot study. These estimates can also be used in official statistics. This “present population” is also called day population oror daytime population is based on registered residents who spend most nights in that municipality.
By Eurostat as well, mobile phone call data are now considered to be a vital source for the compilation and enhancement of official statistics on tourism, economy, environment, crowd flows and mobility. This is also what has spurred Statistics Netherlands (CBS) to research the use of aggregated anonymised mobile call data for statistics since 2009. CBS is no exception in this regard. Around the world, organisations such as universities, national statistical offices, but also commercial organisations are developing methods and software for the analysis and processing of telephony data for the purpose of statistics, science and policy development.
Raw and insufficiently aggregated mobile network data or traffic data are extremely privacy sensitive. During setting the processing method, not only existing legislation requiring the data to be anonymised was taken into account. A series of specific risks were taken into consideration in advance and preconditions have been drawn up for the research to be carried out. These are discussed in chapters 3 and 4. They describe how statistical security measures have been integrated into the design of the method prior to the research. Applying only one measure is not sufficient. Therefore, a combination of several measures as a whole provides the statistical security to ensure privacy.
This paper is an introduction to the technical report . The report describes the mathematical methods that are applied at the Mobile Network Operator (MNO). As a statistical institute, CBS safeguards independence, continuity (guarantee of delivery), international comparability, quality frameworks and transparency of the algorithms. In other words, the algorithms and methods on which the statistics are based must be 100% explainable and transparent. The reports of the study are therefore also made public. This pilot study was completed in 2019.
This paper is made up of several chapters. Chapter 2 briefly discusses the basic principles of the method and roughly describes the processing process. In chapter 3 all the steps of the process are described in detail. Finally, chapter 4 discusses the privacy aspects of the method on the basis of a number of international publications that refer to individual disclosures from aggregated datasets with origin-destination information.