With the recent introduction of modelling using structured/time series data, I was playing around with web traffic since that’s something I’m interesting in examining, especially in terms of security. The kind of data I would be trying out is Content Security Policy reports and HTTP requests.
a) Identify patterns in the web traffic data to distinguish from “good” user behaviour to “bad”
b) Given two web traffic data streams for a user, is it possible to apply machine learning so a model can distinguish between the two streams of data?
I’m wondering if someone has similar experience with something like this that may have able to offer some pointers with something like this… and if there’s other interesting things one can do with web traffic data!