Automated data pipeline (FlowKit) in Mozambique for disaster management and monitoring of displacements
Telecommunications regulator; MNOs (Mcel, Vodacom and Movitel); disaster management agency; Digital Impact Alliance (DIAL); Flowminder Foundation.
The Government of Mozambique, through the telecoms regulator ARECOM and the disaster management agency Instituto Nacional de Gestao de Calamidades (INGC), have partnered with the Digital Impact Alliance (DIAL) and Flowminder Foundation to process de-identified mobile data from three MNOs (Mcel, Vodacom and Movitel) using a semi-automated software tool in order to detect disaster-driven displacements and returns, providing the results in a practical and timely manner. This is being done by identifying changes in the main location of residence of phone users that coincide with the time and location of a sudden onset disaster. INGC will be able to use the system to follow the population movements from the time that early warning messages are first received through the displacement period and beyond to the response and recovery phases. Past mobile data is being analysed to assess the sensitivity and accuracy of the displacement monitoring system for future use, by calibrating it with data from the time and location of the 2019 cyclones Desmond, Kenneth and Idai. As a first step in the process, ARECOM and Flowminder are establishing the data pipeline with daily transfers of de-identified CDR data. Importantly, no personally identifiable data will be used, and the project will only output aggregated data. Access to data will also be restricted to project partners authorised by ARECOM, and will be controlled using FlowKit, Flowminder's open-source toolkit for the management of secure data access and authorisation.
Key steps taken for developing the institutional framework and analytical pipeline
Legal frameworks and partnership agreements in place.
Installation of dedicated server(s) at the regulator premises
Implementing a protocol for pseudonymisation of CDR data
Procuring hardware suitable for the storage of the data
Transferring historical data
Setting up a remote connection for running code and transferring anonymous aggregated results
Areas of improvement and challenges
Countries where pipeline is established - generalising the approach of software suite to different countries and use cases