2.5 million customer reviews and product metadata entries for the Video Games category on Amazon. The collection includes natural language review text, numerical star ratings, and temporal metadata for items sold between 1996 and 2018.
Use Cases
- Train a sentiment analysis model using the `reviewText` column to predict the `overall` rating
- Build a recommendation engine by mapping user-item interactions via the `reviewerID` and `asin` columns
- Analyze consumer behavior over time by grouping `overall` scores by the `unixReviewTime` timestamp
- Perform topic modeling on the `reviewText` field to identify common product defects or praised features
Strengths
- Contains 2.5 million reviews with associated `reviewText` and `summary` strings
- Includes `overall` star ratings on a 1-5 scale for every entry
- Features `reviewerID` and `asin` columns for identifying unique users and products
- Provides `unixReviewTime` and `reviewTime` fields for temporal trend analysis