Canonical components of the canonical analysis of two random vectors are collinear with principal components of a PCA of the multidimensional linear regression function of one vector with respect to the other. In the context of streaming data, we estimate online in parallel this regression function and components of a canonical correlation analysis, taking into account at each step a mini-batch of current data or all the data up to the current step to have a faster convergence, and using extended Oja processes. We extend this approach to generalized canonical correlation analysis and deal with the cases of streaming factorial correspondence analysis, multiple correspondence analysis and factorial discriminant analysis.