Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

When using MPD to produce official statistics, it is important to consider the following:

  • The availability and quality of socio-demographic information. Often, variables such as gender and age are not collected by MNOs or are not made accessible for statistical purposes. Even if such data is available, it tends to be highly unreliable, as the person that a mobile phone number is registered to is often not the same as the person using the mobile phone (e.g. phones of children being registered to their parents).
  • Another possible limitation can be the completeness and accuracy of the data as they depend on the type of devices (i.e., new generations' mobile phones can provide more data than the old generation or obsolete mobile phones).
  • The location is almost always the position of the mobile network cell that the subscriber is connected to. The time gaps between records define the density of the data

    . This issue can largely be overcome by only taking the first and last event in a specific location and removing the in-between data.

  • It is crucial to obtain the data from all the MNOs existing in the country since if a person changes his/her mobile operator and keeps the device, it will be possible to trace further his/her activities.
  • MPD can be difficult to access from MNOs due to privacy and other legal aspects. MPD includes highly sensitive information about the whereabouts of the subscribers (which can present ethical issues) or are difficult to access for commercial confidentiality reasons.
  • When checking for the quality of geographical coverage, the network's cell level is the minimum level of analysis. For more precise analysis, it is necessary to implement various probability models that take into account factors such as road networks, building density, type of land use, etc. Additionally, the uneven distribution of mobile antennas throughout the country generally reflects the location of the population and transport infrastructure. The accuracy of passive positioning is, therefore, greater in more densely populated areas or areas with denser networks of roads, with accuracy being lower in more sparsely populated areas. So, it might be a challenge as the nearest cell tower does not always precisely capture users' locations

    Hughes, C., Zagheni, E., Abel, G. J., Sorichetta, A., Wi'sniowski, A., Weber, I., & Tatem, A. J. (2016). Inferring migrations: traditional methods and new approaches based on mobile phone, social media, and other big data: feasibility study on inferring (labour) mobility and migration in the European Union from big data and social media data.

    .

  • Geographic accuracy (or the so-called "tossing") means that even if individuals are in one place, their phone may connect to different antennas and swap between (toss) them automatically, depending on the antennas' workloads. It is possible to raise the geographical accuracy of the passive positioning data with the cooperation of operators, e.g., dividing cells into sectors or using better positioning techniques for data collection. Additionally, it is possible to use models to assign a user to one cell instead of both.
  • Coping with under- or over-coverage. There is a discrepancy between the target population and mobile phone users. MPD captures only the subscribers, whilst the target population includes all the individuals who reside in the country. Certain people do not have mobile phones, and others have multiple phone numbers. Additionally, there are phone numbers that are not assigned to people but rather companies or machines.