Quantopian at present does not appear to offer reasonable access to past values for fundamental data that spans a couple of years. I keep getting timeout errors as I'm trying to implement the Quantitative Value algorithm. On the verge of giving up.
There are a number of possible fixes for my situation:
- A date argument for
- More efficient access of data points via pipeline. I don't need all data between t0 and t1 to get the values of a factor at t0 and t1.
- Support for multiple pipelines in an algorithm, so that I might at least spread the algorithm's execution over multiple days.
- Generally faster access to fundamental data, right now this is abysmally slow.
- An extended initialization period for an algorithm, so that I may cache some of the values I need for algorithm warm up, such that I can then roll in new values as the algorithm continues to execute. This will of course take up quite a bit of memory in the context object, but will let an algorithm to start executing in live/paper trading right away.
If anyone might be able to offer suggestions, I would love to be able to move forward.