WPI Working area English (en) français (fr)

From ESSnet Big Data
Jump to: navigation, search

The WPI working area contains links to drafts, lists and checklists, and similar documents to prepare for meetings, reports, deliverables for WPI Mobile networks data.


Literature review template

The literature review template can be used by all ESSnet Big Data workpackages to provide WPK Methodology and quality with information on literature (papers, articles, ...) relevant for a particular workpackage. This will be used as input for the WPK literature overview deliverables (K2 and K12).

WP leaders are requested to distribute the link to the template within their workpackage, so participants can provide WPK with information on relevant literature, with a special emphasis on quality indicators for big data statistics (see also the WPI documentation overview page with referenced articles and papers!).

Download the template here, complete it and mail it to WPK (Alex Kowarik (alexander.kowarik@statistik.gv.at) and Jacek Maslankowski (j.maslankowski@stat.gov.pl)).

PLEASE NOTE THESE ARE NOT FINAL DOCUMENTS. THEY REFLECT WORK IN PROGRESS OPENLY SHARED.

WPI on mobile networks data

WPI on mobile network data is structured in parallel, though connected, research tracks. The ultimate goal of this work package is to design an end-to-end statistical production process using mobile phone data. In this line pilots are understood to tackle the many facets of this process and not to show the potentiality of these data for statistical purposes, something which has already been proven in many works in the field (see Documentation).

The different research tracks are thought to deal with concrete aspects of the statistical production process. The research is designed in a modular fashion to avoid the work being blocked due to obstacles in the development of the work. This is especially referring to the remarkable difficulties in accessing mobile phone data by NSIs both for the research and the production of official statistics.

The process under design is in full agreement with the Reference Methodological Framework promoted by Eurostat. Schematically we can represent the framework under construction and the main elements under study as the following figure:

MPD Process

We introduce the different tracks according to this view of the process:

T1. Access

The access to mobile phone data for NSIs in the ESS is still an unsolved question. According to the position paper by the ESS on access to privately held data which are of public interest, we seek partnership models between NSIs and Mobile Network Operators (MNOs) hopefully providing a win-win agreement so that both the ESS (hence society in general) and MNOs see their mutual public and private interest fulfilled thus promoting  the goals of the Digital Single Market Strategy in the field of the data economy fostered by the European Union.

This track entails contacts and joint work between NSIs and MNOs. This is ongoing in the different countries of this work package with diverse situations per country and per MNO.

T2. Synthetic data

This track is devoted to produce synthetic mobile phone data. Ultimately this is an agent-based simulation in which mobile devices are simulated in a given territorial during a given time period. Synthetic data will be used not only when access to real data is blocked but also for model and quality indicators assessment. We use these data to fine-tune different methodological proposals to be applied on real data under standard production conditions.

  1. General proposal: A framework for mobile network data micro-simulation (PDF.jpg PDF, 112 kB)

T3. Methodology

This track embraces methodological proposals of diverse aspects of the process (from the geolocation of network events to inference models based on aggregate mobile phone data).

  1. Exploratory proposal following the use of hierarchical models with admin data: An inferential framework for official statistics based on mobile network data: Bayesian hierarchical models combining auxiliary information (PDF.jpg PDF, 437 kB)

T4. Information technologies

This track embraces the software development and related IT issues regarding the implementing of the preceding methodological proposals. The goal is to produce a modular software tool embedding the different aspects regarding the use of mobile phone data for the production of official statistics (from the generation of simulated data to the inference upon target populations over the geolocation of network events).

For code see github repository.

T5. Standards/Metadata

This track embraces proposals to adopt standards and to build a system of metadata for the production framework under construction.

T6. Quality

This track embraces the assessment and related proposals of the different quality issues about the statistical process under construction.

T7. Application on real data

This track compiles all preceding proposals and tries to apply all of them to real data to simulate an end-to-end process producing official statistics.

T8. Visualisation

This track compiles specifically those issues related to the visualisation not only of final outputs but also and equally importantly of intermediate production steps (e.g. for the geolocation of network events).

For code see github repository.