Geography Reference
In-Depth Information
Then, the Moran's I local indicator of spatial association (LISA) (Anselin,
1995) was calculated for each cell in every three hours (See Figure 11). The
value of Moran's I has been standardized to lie in [-1, 1]. If the index is larger
than 0, the cell shows positive spatial autocorrelation with its neighbors; while
it indicates the negative spatial autocorrelation of phone call patterns if the
index is smaller than 0. The closer the value is approaching to 1 (or -1), the
stronger the positive (or negative) spatial autocorrelation is. Examining the
spatial structure of LISA in different time periods, we can clearly see that the
spatial autocorrelation patterns of phone calls across all cells are very dynamic
and heterogeneous.
The central region (small cells) shows more diverse patterns than the outer
suburb areas, where most spatially adjacent cells show similar values in the
whole day.
It might reflect the mixture land-use types of urban central areas and
human's convergence and divergence in this place with various phone call
behaviors in different time periods.
To identify a more stable autocorrelation structure, we apply the spatial
statistic test of running 10000 simulations of randomized permutations of
neighboring cells to find the local significant spatial autocorrelation patterns
(Anselin, 1995).
Figure 9. Time-series plot of phone-call net flow among all mobile cells (each line
represents a different mobile cell).
