Figure 3
From: Prediction of employment and unemployment rates from Twitter daily rhythms in the US
The result of the principal component analysis of the population-weighted covariance matrix. (a) Proportion of explained variance for the principal components of the covariance matrix. Only the first four components carry a share of variance significantly greater than zero. (b) Principal component corresponding to the largest principal value. The amplitude of vector components is plotted, each vector coordinate corresponds to an hour ranging from 0 to 23. The vector components are positive from 5am to 8pm, and negative otherwise.