Disrupting data discovery

Mark Grover discusses how Lyft has reduced the time it takes to discover data by 10 times by building its own data portal, Amundsen. Mark gives a demo of Amundsen, leads a deep dive into its architecture, and discusses how it leverages centralized metadata, PageRank, and a comprehensive data graph to achieve its goal. Mark closes with a future roadmap, unsolved problems, and collaboration model.

Before any analysis can begin, a data scientist needs to discover the right data sources to analyze, understand them, and gain trust in them. Unfortunately, data discovery is very inefficient today. Countless hours get lost trying to find the right data to use; the most common way still remains to ask a coworker. Gaining trust in data requires running a bunch of queries—max timestamp, counts per day, count distincts, etc.—that waste time and add unnecessary load on the databases. There’s no clear way to know how to find folks to answer questions about the table. And, worst of all, many times analysis is redone and models are rebuilt because previous work is not discoverable. Mark Grover discusses what a data discovery experience would look like in an ideal world and what Lyft has done to make that possible. Lyft has reduced time spent on data discovery 10 fold because of its data portal, Amundsen. Amundsen is built on three key pillars: An augmented data graph Amundsen uses a graph database under the hood to store relationships between various data assets (tables, dashboards, protobuf events, etc.). What’s unique to Amundsen is that it treats people as a first-class data asset—in other words, there’s a graph node for each person in the organization that connects to other nodes (like tables and dashboards). An intuitive user experience Amundsen runs PageRank using data from access logs to power search ranking, similar to how Google ranks web pages on the internet. Centralized metadata Amundsen gathers metadata from various different sources (Hive, Presto, Airflow, etc.) and exposes it in one central place. The right place to store all this metadata is a work in progress. Mark shares ongoing efforts in this space, including RISELab’s Ground and WeWork’s Marquez projects. Mark gives a demo of Amundsen and its goals, leads a deep dive into Amundsen’s architecture, and explains how it achieves the three design pillars. Mark closes with a future roadmap of the project, what problems remain unsolved, and how we can work together to solve them.