How to model multiple time series samples?

I have time series data for 40,000 customer accounts energy consumptions through 10 year peroid at daily frequency. I'd like to find the customer energy consumptions changes through the time for the whole population.In other words, the annaul energy consumption growth in percentage for the entire population. The customers have different account histories. Some of them have 10 years account, some of them have less than 10 years.

I am planning to use linear regression to find each customers growth rate with each customers daily consumption. With each customer sample, I can get the all statistic data such as R value, slope, rediduals. But I am not sure how to get the growth rate for the population and r values from each customer sample.

Thanks for the help in advance.

Last edited: