I have refined the analysis in the main entry to include more glitches and all the auxiliary channels measuring positions, velocities and accelerations, as described in this git issue. The results are contained in the attached text files. These confirm and expand what is already presented in the main entry, that is, a strong correlation of these glitches with the movement of the West End bench and suspension. Correlations with velocity and acceleration sensors are instead much smaller. They tend to point to the motion of parts of the North End building, but plotting the predictor frequencies superimposed to the spectrogram of Hrec shows in fact no convincing correlations. Example in figure 1.