Competing destination for Spatial Interaction models, SpInt/PySal

Competing Destination (CD) models, which are an extension of Spatial Interaction (SI) models, has been around since 1980's and are often used within Economic and Social Sciences . CD models involve the analysis of flows from an origin to a destination similarly to traditional SI models, however, CD includes the ‘Competing Destination’ term also known as ‘Accessibility’ term, which accounts for the spatial-structure effect of the flows from behavioural perspective. Although the SI models are established within Python and R modules already (SpInt, simR), specification for deriving Accessibility term for CD estimation is missing. This project aims to fill this gap by developing a Competing Destination class that will include the computation of the accessibility term and will be binded to the existing SpInt module. The main challenge here is the scaling of the accessibility computation for large datasets. This is an important aspect of the project as flow datasets often have hundred thousands and even millions records. This project will extend the use of the SpInt module and PySAL library for high level analysis of spatial flows.