Crisp provides sample data sets that allow you to experiment with building models using data that mimics the real-world sales and supply data of a consumer packaged goods (CPG) brand. 

Any authenticated Google user can access our public data sets and objects using the following links.

Note: While Crisp pays for the primary storage of these data sets, the Google account accessing them will be responsible for subsequent downloading and processing costs. 

BigQuery data sets
Cloud Storage buckets

Data transformation details

While this data is representative of real data, any identifying data (such as product, brand, supplier, retailer etc.) has been consistently obfuscated and masked to prevent mapping back to any real world entity, including slightly randomizing select values (such as sales amounts and quantities). Crisp identifiers in the data have also undergone consistent transformation. A Crisp ID that appears in these data sets will not map back to the same entity on the Crisp platform.

While physical addresses have been masked, certain geographical aspects of the data sets remain intact. For instance, postal codes and latitudes and longitudes remain intact with only coarse transformations applied to limit their specificity (e.g., latitudes and longitudes are truncated to two decimal places).