data on flavor of 86 single malt scotch whiskies

Dataset of ~50K Kindle book reviews, all containing a description from that book.

Movies dataset from 2000 to 2018 used to predict 2019 Academy Awards

Kiva is a non-profit organization with a mission to connect people through lending to alleviate poverty.

Source:

Loans from a snapshot at build.kiva.org. A Python script was used to process the JSON files that were downloaded from the snapshot, resulting in a .csv with most pertinent data fields.

Learn more in this blog post.

dataset of web pages from stumbleupon

Description List of properties for sale in Corvallis, Oregon, USA.

Source Dataset from Redfin.

Dataset that classifies handwritten digits using data taken from the MNIST database. Source: kaggle.
Originally from Yann LeCun MNIST database.

Model with word counts for about 47,000 text documents, of which roughly 32,000 are novels, 7,500 are supreme court opinions, and 7,500 are webpages from universities. The features are word counts for the 3000 top words by TF*IDF, with stopwords removed.

The novels are from Project GutenbergFDSys, the supreme court cases from FDSys, and the web pages from web->KB.

The dataset on the distinction between good and bad connections (intrusions was part of the data created for The Third International Knowledge Discovery and Data Mining Tools Competition, which was held in conjunction with KDD-99 The Fifth International Conference on Knowledge Discovery and Data Mining.