Download as PDF (English) | Télécharger le PDF (en français)
Executive summary
Key observations for the period 01 March 2024 to 28 February 2025
-
Of the communal sections for which we can provide monthly population estimates (273), 8.1% (22) exhibited large overall increases (> 1,000) while 4.8% (13) saw large overall decreases (< -1,000).
-
The largest changes were within the Port-au-Prince Metropolitan Area (ZMPAP), likely driven by insecurity due to gang violence, with the largest decreases in 1re Section St. Martin and in 1re Section Turgeau (suggesting significant insecurity), the largest increase in 5e Section Bellevue Chardonniere, and high levels of mobility between these three sections.
-
Outside West, there were large increases in 1re Section Bourdet (Les Cayes) in South, as well as 2e Section Haut du Cap (Cap Haitien) in North and 6eme Section Charrette in Artibonite.
-
-
10.6% (29) of communal sections with recent estimates had fluctuating populations
-
Fluctuations can show changes in population likely to correspond to displacements and returns following armed attacks (e.g. in Artibonite), or indicate a data issue where there are very large and sudden changes (e.g. in Nippes and North-West).
-
-
At the department level, West saw the largest estimated change, shrinking by about 37,500 people, mostly occurring in the first half of the study period, then levelling out.
- Conversely, Centre, North, and North-East exhibited relatively steady growth. Elsewhere, South, South-East, Nippes, all showed overall growth, with steady increases early in the reporting period before leveling off, or following a downward trend, in the latter half.
How to use this report
We provide population estimates that take into account population movements within Haiti, per month and per communal section. The current crisis has led to large population movements, particularly of internally displaced people (e.g. IOM round 9), which can result in population statistics becoming outdated over a few months, and therefore impairing preparedness efforts (e.g. hazard population exposure estimation, needs assessment, service and contingency planning).
Indicators presented
Modelled change (trend): the predicted population change for the reporting period from all 12 monthly estimates, which smoothes out monthly variations (linear regression).
Estimated change: the difference in the estimated population between the first and last months of the 12 month reporting period, which does not reflect any monthly variation within.
For communal sections with fewer than 10 months of resident estimates during the reporting period, we do not make a trend classification (insufficient data). For monthly population estimates at the department level, we carry forward the latest estimate to fill missing months in a communal section so that the aggregated changes at the department level are only based on observed CDR-derived mobility.
For communal sections with fast changing populations, we categorised them as ‘increasing’, ‘decreasing’; ‘stable’, when there is little to no change; or as ‘fluctuating’ when large month-to-month changes prevented us from detecting a trend.
Data used
- This report uses our v4.0 dataset (documentation, release notes), available at https://haiti.mobility-dashboard.org.
- Monthly estimates of population per communal section (the “de facto” population) - derived from anonymous mobile phone data, weighted using survey data but unadjusted for population change due to births, deaths, immigration and emigration - only considering internal mobility in Haiti.
Visit our Haiti Mobility Data Platform.
About this report
Authors & contributors
This report was authored by the Flowminder Foundation, by Véronique Lefebvre and Zachary Strain-Fajth, with the contribution of Robert Eyre, John Roberts, Roland Hosner and Sophie Delaporte.
Zachary Strain-Fajth produced and analysed the mobility statistics and co-wrote the report; Véronique Lefebvre directed the analysis, interpreted the mobility statistics and co-wrote the report; Robert Eyre wrote the notebooks that generate the indicators; John Roberts developed the report automation pipeline; Roland Hosner developed and applied the bias-adjustment and scaling method, and Sophie Delaporte supported with information product design, report review, translation and data visualisation.
Acknowledgements
This study was made possible thanks to the anonymised mobile phone usage data provided by Digicel Haiti, which are aggregated by Flowminder via FlowKit to provide statistics.
This work has been made possible thanks to funding from the UK 's Foreign, Commonwealth and Development Office (FCDO)
Data privacy & governance
No personal data, such as an individual’s identity, demographics, location, contacts or movements, is made available to the government or any other third party at any time. All results produced by Flowminder are aggregated results (for example, subscriber density in a given municipality), which means that they do not contain any information about individual subscribers.
This data is fully anonymised. This approach complies with the European Union’s General Data Protection Regulation (EU GDPR2016/679). Data is processed on a server installed behind the mobile network operator’s firewall in Haiti, and no personal data eaves the operator’s premises.
Data considerations
The estimates shown are our best current assessment of movements. However, there are a number of uncertainties. The information should be interpreted together with other available evidence.
Data used
This report uses our v4.0 dataset (documentation, release notes), available at https://haiti.mobility-dashboard.org.