These datasets are gathered from Coffee Quality Institute (CQI) in January, 2018. I'm not the Owner of the Datasets, nor scrapping was performed by me. It was done in this GitHub's repo (kudos for the author), see there for further details.
Three CSV files are provided:
An Arabica coffee pre-cleaned dataset;
A Robusta coffee pre-cleaned dataset;
A dataset constructed through a merging of the datasets.
The file names indicates the above datasets clearly.
As explained in the repo, the datasets have reviews from specialized reviewers for both coffees: arabica and robusta. The below information is provided in each dataset.
There is one related dataset here in Kaggle, please check here. It's pretty much similar to the datasets presented here, but without Robusta coffee data.