Crisis Preparedness

Well thought-through and robust preparations for crises should help respond more effectively and ultimately bring stronger benefits for the people impacted.

CDR data can support these preparations and inform crisis planning in a number of ways. We can analyse CDR data to provide insights into the population prior to a crisis, inform predictions about the impacts of a crisis, or for use to simulate crises.

Dynamic mapping using CDR data can provide an up-to-date and nuanced picture of how the geographic distribution of the population varies over time. This may be short-term variation in the number of people visiting different locations (e.g. at different times of day and night, different days of the week) or the longer-term variation in the number of people residing in an area (e.g. seasonally). This information can help crisis management planners to understand how many people may be impacted by a crisis and how this might be affected by the timing (e.g. weekday vs weekend or between seasons). This is especially useful in the case of a forecastable crisis, such as a hurricane or cyclone, where the impacted area and timing can be projected.

We can also analyse CDR data to predict internal displacements in the event of a crisis. Flowminder developed predictive models of crisis-driven displacement for three natural disasters: the 2010 earthquake and 2016 Hurricane Matthew in Haiti, and 2015 earthquake in Nepal. These models showed that the locations visited by an individual prior to displacement as well as the ‘home location’ of social contacts were important predictors of displacement distance and location; displaced people will, if possible, relocate to or stay in areas they have connections with and mobile usage data can identify these areas. Individuals with localised travel patterns and social contacts were more likely to be displaced in the vicinity of their pre-disaster residence, regardless of disaster intensity and damage to property at the location, compared with those with more dispersed travel patterns and social contacts. Of individuals moving long distances,  60-70% were displaced to locations they had previously been observed to visit.

All of this information can support the development of crisis preparedness tools to simulate crises. Using indicators of the geographic distribution of the population and predictive models of internal displacement and resettlement calibrated on past events, it would be possible to forecast the potential outcomes of crisis scenarios and use this information to guide decision-making and improve preparedness. In particular, simulations could be used to assess the resilience of different locations or areas likely to receive large numbers of displaced people to inform the distribution of resources and investment.

The use of CDR data to support crisis preparedness has the additional benefit of establishing the circumstances which best facilitate the continuous use of this data source in the event of future humanitarian crises. Negotiating the necessary access and setting up the data access will facilitate the rapid production of mobility indicators in response to a crisis and the data and analyses carried out for crisis preparedness will help establish the baselines against which disruption associated with the crisis will be measured.