Conclusions

• OAC can be used simply as a contextual reference
• OA level interaction data have to be treated with caution, but are not useless
• OAC can enhance the OA level interaction data
• Data can be aggregated on the basis of OAC as well as on the basis of geography
• Spatial information can be retained by combining OAC aggregation with distance-travelled information

Notes:

This has explored some of the ways in which OAC and OA interaction data can be combined; OAC can enhance the OA level data

It would be possible to use ward or district level geo-demographic coding as an alternative; this would have the advantage that the numbers in each flow would be larger. However, note the trade off between spatial resolution and geodemographic rigour. If only one zone was used, then we could only have one area type! As the number of areas increases, so we can support multiple clusters. But this does not continue ad infinitum: at some level, there are too few people in an area to be able to classify consistently. The appropriate level probably depends on the phenomena being studied. For many purposes, districts are too large and heterogeneous to wish to classify them on the basis of the people living in them. Wards may well be more fruitful.