You need less data
Often data seems to be big. But what is useful is a much smaller subset due to mathematical and/or data quality issues.
On the other hand, You need more data.
References:
- https://statmodeling.stat.columbia.edu/2005/07/31/n_is_never_larg/