hi @RishiPSingh_twitter, the repo is designed around the pandas DataFrame format, and the notebooks are ideally replicable offline, so websockets for dynamic data would be secondary in importance. If you optionally include it though, it would be very instructive to many interested in online analysis and trading. I would be willing to write a tutorial, say, to demonstrate Kalman filtering.
As for websocket preference, Tornado http://www.tornadoweb.org/en/stable/websocket.html is extremely reliable, performant, and the code is well-written. In fact, it is embedded into Jupyter.