Numerai Sunshine dataset v4.1 as of May 25, 2023, divided into train and validation and both in .parquet and .csv format. This dataset is obfuscated global equity data used to create predictive models for the daily Numerai data science tournament.

See Numerai's Terms & Conditions for data license.

What we have here

This is a dataset released in Sep 19, 2020 by the Numerai for its tournament.

The training data as well as the validation data here have two target types ('kazutsugi -- old target, 'nomi' -- new target).

By using this version of the training data, you can use the old target 'kazutsugi' as well to train and validate your model, which may be advantageous because the recent training data only contain the new target ('nomi').

Note that originally the data were provided by the Numerai by this forum: New Target Nomi Release. I just kept it here for those who are interested in using the old target 'kazutsugi' for their modeling within the Kaggle environment.

For those who seek for the latest train and tournament data, use the following notebooks.

• [Numerai] get tournament data
• [Numerai] get train data