Solar power production in MW recorded per every 4 seconds starting from 01/08/2019 in Australia.
From the website:
-----
This dataset contains a single very long daily time series representing the solar power production in MW recorded per every 4 seconds starting from 01/08/2019. It was downloaded from the Australian Energy Market Operator (AEMO) online platform. The length of this time series is 7397222.
-----
Here is the dataset curated by the Monash Time Series Forecasting Repository. It is not clear which were the preprocessing steps and how did they
acquired the data from the original website (https://aemo.com.au/ and http://www.nemweb.com.au/).
There are 4 columns:
id_series: The id of the time series.
date: The date of the time series in the format "%Y-%m-%d".
time_step: The time step on the time series.
value_0: The values of the time series, which will be used for the forecasting task.
Preprocessing:
1 - Renamed columns 'series_name' and 'series_value' to 'id_series' and 'value_0'.
2 - Exploded the 'value' column.
3 - Created 'time_step' column from the exploded data.
4 - Cretead 'date' column from the 'start_timestamp' and 'time_step' column, by offseting the 'start_timestamp' by 4 secods * time_step.
5 - Dropped 'start_timestamp' column. Defined 'id_series' as 'category' and casted 'value_0' to float.