When using multiple data streams of this nature, there will inevitably be multicollinearity.
The paper by Prestige et al. looks at relationships between Mobility and CoMix and can a priori indicate where strong relationships lie, including non-linear relationships between the data sources.
Could use VIF or equivalent plots. Consider using regularisation techniques within modelling to promote stability of coefficients. I anticipate this to be an issue determining relative contributions when using CoMix and Mobility in the same model
When using multiple data streams of this nature, there will inevitably be multicollinearity.
The paper by Prestige et al. looks at relationships between Mobility and CoMix and can a priori indicate where strong relationships lie, including non-linear relationships between the data sources.
Could use VIF or equivalent plots. Consider using regularisation techniques within modelling to promote stability of coefficients. I anticipate this to be an issue determining relative contributions when using CoMix and Mobility in the same model