Feature Store Ranking

Jan 2020 Rank Dec 2019 Rank System Jan 2020 Score Dec 2019 Score
1Hopsworks 7.5
2Michelangelo 3.0
3Feast 2.5
4Bighead/Zipline 1.5
5Netflix Metaflow 0.5
6Conde Nast 0.4
7Facebook FBLearner 0.3
8PInterest Galaxy 0.25
9Comcast 0.2
10Twitter 0.1

Ranking Methodology

The Feature Store Ranking is a list of feature store systems for ML ranked by their current popularity. The popularity of a feature store is measured using the following criteria:

  • Frequency of links/mentions on websites, measured as number of results in search engines queries. Google allows us to count searches for <system> together with “feature Store”, e.g., “Michelangelo” and “feature store”.
  • Search interest. The frequency of searches using Google Trends.
  • Technical discussion on Stack Overflow. The number of related questions and the number of interested users.
  • Interest in Twitter. The number of tweets that reference the system.

We calculate normalize and average the different criteria when calculating the popularity of a system. The score can be seen as a measurement for the system’s popularity (3X score, means roughly 3X times the interest in a system). We are not able to count the number of installations.