Accessibility mapping as a tool for measuring rural deprivation

[ Contents ]

Data requirements

Almost any type of geographical feature which might influence travel time can be incorporated into an accessibility model. However, so many assumptions have to be made about how quickly people can move through the landscape, that it is counter productive to use too many datasets. At best the model can only give approximations and comparisons and should not be treated as significantly precise or accurate.

In the case study given here, it is assumed that travel will primarily be along roads and that travel times will be influenced by the quality of the roads and the steepness of the land. In order to produce the accessibility map, the following data sets will be needed:

Road map, showing categories of road

• DEM or a derived slope map of the region
• Target features: places of interest that people will travel to or from. In this case study there are two types of targets:
  • Settlements: location of centres of population, whose deprivation in terms of access to services is to be determined by the model
  • Services: the location of features such as schools, hospitals, Village Knowledge Centres (VKC) and anything else of interest.

Additional data that might be added to the model are landcover, if a significant amount of travel will be off road, and rivers, if these are used for transport or are significant barriers to land transport i.e. they could both increase or reduce travel times. If there are definite barriers to travel, such as large bodies of water or national boundaries, these should be added to the model. None of these additional features are considered in this case study, which simply intends to show the principles of the technique.

Figure 3 shows the basic distribution of the datasets in the case study region (this is an area around the borders of Guatemala, El Salvador and Honduras. The villages, roads and slope are true features, the services targets are hypothetical.)

Figure 3 Basic data sets used for this case study. These are based on data from Central America but the service targets (schools, hospitals and VKC) are hypothetical.

Source: CIAT 2006

Data type and projection

Vector data for roads and targets can be used, but the accessibility model is generated as an raster surface map. All vector data sets will have to be converted to rasters. All datasets must be projected on a common equal-area projection for the calculated travel times to be anything close to real times.

The resolution of the raster is important as it relates directly to travel time. The times used are those required to move either directly or diagonally across one cell of the raster at any location. How the resolution is incorporated into the calculations will vary depending on the algorithm of the geographical information system (GIS) used to build the model.

^^top [ Contents ]