Prepositioning relief: Using a random walk model to predict the distribution of IDPs in Nigeria


Nov 07 2017 Print

Nov 07 2017

Random walk is a mathematical modelling technique used in many scientific fields to model seemingly random behavior. Displacement in conflict scenarios, in contrast to displacement induced by rapid-onset natural disaster event, appears to be “random” in that it does not follow a consistent or clearly predictable pattern. IOM piloted use of the random walk technique to conceptualize and model the geographic distribution of future displacement in Nigeria, using displacement data collected through DTM since 2011. In an initial pilot, the model proved effective in forecasting displacement trends in the areas monitored by DTM, with a maximum error of 10% between output of the simulation and the actual data collected. In practice, this means the model was able to anticipate with a high level of accuracy, based on earlier rounds of data collection, the movements of approximately 1.8 million people between rounds of data collection. This technique may assist future operations, including informing the planning and prepositioning of relief and services. This project is in its pilot phase and is currently under development. Scroll down for more details.


For more information please contact:

Eduardo Zambrano
Statistics and Reporting Consultant