Geospatial Social Science and Human Activities

a. Urban hot spot from tweets

• Mining spatio-temporal pattern with geo-tagged tweets
• Building machine learning tools for human mobility modeling and prediction

Figure. Tweet densities in four US college cities

b. Events detection from tweets

• Using Spatial-Temporal Density-Based Spatial Clustering of Applications with Noise (ST-DBSCAN) to spatially-temporally cluster the tweets
• Modeling potential topics by the Latent Dirichlet Allocation (LDA) algorithm.

Figure. Spatial-temporal clustering results for the tweets in West Lafayette on 21 January 2014 when the gunshot occurred: (a) cluster centers and the most significant cluster; and (b) distribution of the tweets in the most significant cluster. (d1 = 200 m, d2 = 60 min, minPts = 10).

c. Spatial distribution of human sentiments

• Using Global Moran’s I to measure overall spatial autocorrelation based on tweet locations and polarity.

Figure. Positive (hot) and negative (cold) human sentiments in West Lafayette, IN.