Figure 3 | Palgrave Communications

Figure 3

From: Race, religion and the city: twitter word frequency patterns reveal dominant demographic dimensions in the United States

Figure 3

Census data with most correlation for each component displayed on a map (a–e) and scatter plots of these data series with county weights from corresponding singular vectors (f–j). Lines representing a symmetric relationship (Pearson correlation) are drawn on scatter plots (g) and (h). On scatter plots (f), (i) and (j), horizontal and vertical lines correspond to the best segmentation when testing for a Boolean relationship between variables. The points between the dashed lines were not taken into account when calculating the test statistics for the sparseness of each segment.

Back to article page