Time series for network matrices?

Let's say I have two people. Each of whom I create a network matrix for daily for some of their daily habits $(i.e. A, B, C, D)$. I do this for, let's say $60$ days.

Two People and their daily matrices http://www.chrisjvargo.com/Matrix%20Time%20Series.png

I didn't bother to change the values in the matrices, but let's assume they vary from day to day. At the end of the $60$ days, I want to see if Person 1's matrix of daily habits influenced Person 2's matrix. If this wasn't a matrix, I'd do something like an ARIMA time series. Is there a way to process matrices and see if Person 1 predicts Person 2 or vice versa? Open to any statistic/package that might get the job done.

• One possibility: you might stack the values into a vector and use vector autoregressive models (VAR models). – Glen_b Nov 13 '13 at 22:24
• Link doesn’t work – samplesize1 Sep 22 '18 at 16:16