Book page

Mobile Network Operator data for Official Statistics (MNOdata4OS)

0 comments

Context

The European Statistical System (ESS) is undertaking several innovation efforts in order to produce new statistics, with improved timeliness and richer information content, for the benefits of the whole society. As part of this transformative process, the ESS and its members are working towards the (re)use in the statistical production processes of new sources of data, including data generated and held by the private sector. The location data that are routinely collected by Mobile Network Operators (MNO) – or “MNO data” for short – are one of the most appealing candidate sources for (re)use in official stastitics, but also one of the most challenging.

During the last decade several National Statistical Institutes (NSI) in Europe and worldwide have been experimenting with MNO data through case-studies and research projects spanning various statistical domains, e.g.,  tourism and demography. Such explorative activities have clearly demonstrated the capability of such data to capture and trace changes in human mobility patterns at scale, thus enabling the production of statistics delivering a dynamic view of population presence and mobility. In a few cases the collaboration with MNOs has led some NSI to publish Experimental Statistics (see here for a definition) based on MNO data. The covid19 pandemic has highlighted the critical and urgent need for producing timely figures about human presence and mobility and therefore has further accelerated the interest in statistics based on MNO data.

In order to move beyond explorative activities, research projects, one-off case studies and Experimental Statistics towards regular production of Official Statistics based on MNO data, several challenges must be addressed.

  • Data protection. The (re)use of MNO data for statistical production must follow a clear privacy-first approach: the collective benefit of public and freely available official statistics must not come at the expenses of individual privacy rights. As the processing of MNO data qualifies as processing of personal data, the strongest possible technical and organisational measures must be put in place in order to guaranteee the highest possible level of data protection. This is not only a legal obligation but also an ethical imperative! 
  • Enabling legislation and sustainable data access models. The (re)use of MNO data for official statistics must be based on a consistent legislative framework enabling the development of sustainable partnership models between statistical authorities and MNOs.
  • Methodologies. The process of transforming large streams of "raw" data points into a parsimonious set of aggregate statistical data with adequate levels of quality requires the development of new, quite involved statistical methodologies. 

Eurostat is actively engaged to create solutions at the European level for these issues. The relevant initiatives and activities conducted by Eurostat since 2018 are listed below. 

Presentations

Recent and ongoing activities by Eurostat 

  • Legislation and data access. 
  • Data protection. Eurostat is promoting the adoption of Privacy Enhancing Technologies (PET) in Official Statistics, specifically (but not exclusively) for what concerns the (re)use of privately-held data for statistical purposes. For an up-to-date overview of other activities by Eurostat in the field of PET refer to the PET4OS page. 
    • Eurostat-Cybernetica project (2021). This project developed a proof-of-concept solution for the secure processing of longitudinal MNO data.
  • Methodological development. Eurostat is conducting multiple activities in the field of methodological development for MNO data in collaboration with external partners. The list of recent activities includes the following items: 

Papers

Research activities conducted internally in Eurostat and smaller collaborations with external partners led to the following publications

Past projects and activities

  • Eurostat-KidoDynamics project (2021-2023). This project enabled Eurostat to explore macroscopic aspects of real-world MNO data and test some preliminary ideas and methodological solutions on real data. The project was instrumental for Eurostat to gain insight and build knowledge on several the methodological and practical aspects of MNO data processing. During the course of the project Eurostat had no direct access to MNO data but only to the (anonymised) aggregate results produced with queries and algorithms developed by Eurostat and implemented by Kido Dynamics - a company specialised in mobile data analytics. The project was concluded in February 2023. Based on the results of this successful collaboration, a joint paper between Eurostat and Kido Dynamics is currently under preparation.   
  • Eurostat-Proximus project (2019-2020) focusing on the comparison of different geo-location methods for the estimation of present population. The successful collaboration with proximus resulted in the joint scientific  article "Towards a methodological framework for estimating present population density from mobile network operator data" (2020).  
  • ESSnet Big Data I (2016-2018) and ESSnet Big Data II (2018-2020). Eurostat has funded two research grants for ESS members to explore the applicability of non-traditional data sources for official statistics. Each of these two projects had a work-package focusing on MNO data. All their deliverables and results are publicly available here.
  • Feasibility Study on the Use of Mobile Positioning Data for Tourism Statistics, consolidated report 25 June 2014.
Files

Be the first one to comment

Please log in or sign up to comment.