Clarification of Problem Objectives

Power Laws
1

Hello everyone,
I apologize if this question may seem obvious to some, but I have spent the past few days truly trying to understand and explore the data as best as I can. I am participating in the Power Consumption Predicting “branch” of this competition.

Clarification Point #1: (From the prompt)
More than 200 building sites are considered. Three time horizons and time steps are distinguished. The goal is either:

To forecast the consumption for each quarter of an hour over the next 24 hours (96 quarters).
To forecast the consumption for each hour over the next four days (96 hours).
To forecast the consumption for each day over the coming month (30 days).

Does this mean that our forecast will only include 96 “data points” for a particular 15-min model (for example)?
To clarify, I will provide an analogy - is this problem similar to that of stock prediction? Given historic data, forecast the stock price 96 periods into the future (for 15 minute resolution stock price data, for example.) I hope that analogy explains my question.

Clarification Point #2: (From the prompt)
…Participants are not supposed to attempt to get additional external information about the building sites. Also, when forecasting values for a given time period, participants are also not allowed to use given data from the following period (no interpolation is allowed).

So I understand I am not allowed to use data other than that provided - this is clear. But, the final point is unclear to me. Can someone please explain this to me, it seems ambiguous in this context?

Thanks so much in advance, I appreciate any help!

2

Does this mean that our forecast will only include 96 “data points” for a particular 15-min model (for example)?
To clarify, I will provide an analogy - is this problem similar to that of stock prediction? Given historic data, forecast the stock price 96 periods into the future (for 15 minute resolution stock price data, for example.) I hope that analogy explains my question.

My approach has been to use the sample submission file to tell me when to make a prediction. There’s also the submission forecast period file which is meant to help but it seemed easier to me to just to use the sample submission file.

3

Thanks for your reply, I appreciate it! That should help push me in the right direction.
As for my second point, do you perhaps have insight on that? What do they mean by “No data from the following period?”