the weather data comes with predictions. However, they seem to be generated by a perfect oracle, is this intended? Every prediction comes true exactly, far from a realistic scenario. (also, this was different in previous CityLearn challenges)
If this is intended, I’m interested in the reasoning
PS: The submission dashboard says I have 5 submissions remaining. Is this in total or in Phase 1?
Hi @ludwigbald and thank you for your question. Indeed, the weather predictions are ‘perfect’ and it was intended. You are correct that in previous competitions, the predictions were shifts of the true value with some noise added.
In this year’s competition, because the environment lets you have access to those forecasts, doesn’t mean it is useful to solve the problem so your solution may be one that exclude those states.
I’ll refer the submission count question to @dipam to answer.
Please, let me know if you need further clarification.
@ludwigbald : You have 5 submissions per day.
Best of Luck !
Hello @kingsley_nweye ! Is it possible to check if the weather file is from the same location/time as our buildings in stage 1? For me it looks wrong…
Hi @tymur_prorochenko! thanks for your question. The weather file is not from the exact location. Instead, it is the TMY file for the nearest location. Please, do you mind sharing the discrepancies you are noticing between the building and weather data?
If we simplify a lot, pv generation should be very correlated with solar irradiance. However, in our case, solar irradiance does not match to pv power almost at all. Will dm more details.