This paper explores the potential of volunteered geographical information from social media for informing geographical models of behavior, based on a case study of museums in Yorkshire, UK. A spatial interaction model of visitors to 15 museums from 1
79 administrative zones is constructed to test this potential. The main input dataset comprises geo-tagged messages harvested using the Twitter Streaming Application Programming Interface (API), filtered, analyzed and aggregated to allow direct comparison with the models output. Comparison between model output and tweet information allowed the calibration of model parameters to optimize the fit between flows to museums inferred from tweets and flow matrices generated by the spatial interaction model. We conclude that volunteered geographic information from social media sites have great potential for informing geographical models of behavior, especially if the volume of geo-tagged social media messages continues to increase. However, we caution that volunteered geographical information from social media has some major limitations so should be used only as a supplement to more consistent data sources or when official datasets are unavailable.
Iterative proportional fitting (IPF) is a widely used method for spatial microsimulation. The technique results in non-integer weights for individual rows of data. This is problematic for certain applications and has led many researchers to favour co
mbinatorial optimisation approaches such as simulated annealing. An alternative to this is `integerisation of IPF weights: the translation of the continuous weight variable into a discrete number of unique or `cloned individuals. We describe four existing methods of integerisation and present a new one. Our method --- `truncate, replicate, sample (TRS) --- recognises that IPF weights consist of both `replication weights and `conventional weights, the effects of which need to be separated. The procedure consists of three steps: 1) separate replication and conventional weights by truncation; 2) replication of individuals with positive integer weights; and 3) probabilistic sampling. The results, which are reproducible using supplementary code and data published alongside this paper, show that TRS is fast, and more accurate than alternative approaches to integerisation.
