Fitting a linear model to bacteria population change