Geography Reference
In-Depth Information
Fig. 6 Socio-demographic variables for tweet hotspots
The visual exploration of demographic variables serves the purpose of indicating
candidate variables for an ordinary least squares (OLS) regression model with the
purpose to explain tweet numbers through these variables. The OLS regression
model of the variables show in Fig.
6
lead only to a low model fit, with R
2
being 21.
That means that there is no strong statistical relationship between the selected
socio-demographic variables and tweet counts. The result does not suggest that a
spatially-explicit regression model is required as the variables have weak explan-
atory power on a global level. We conclude that socio-demographic variables do
not directly explain tweet counts in London
s wards.
The weak results of the statistical analysis could be influenced by the level of
detail of the socio-demographic variables. The age groups are not sufficiently
differentiated in these variables, as all inhabitants between 16 and 64 years are
'
