An activity schema is the fastest way to work with data
One transformation per source
An activity schema is created by a set of 20-30 line SQL transformations directly from raw data sources. Each defines a single business concept — orders, page views, support tickets, etc.
An activity schema only requires one model per concept — with no foreign keys to each other, no dependencies on other models, and no pre-built aggregations.
Builds one activity table
A single time-series table that contains all the data you need.
All downstream plots, tables, materialized views, etc used for BI are queried directly from this single table, with no other dependencies.
Query anything for analysis and BI
Any activity can be directly combined with any other in a query — without requiring foreign keys to join. For the first time, all data sources can be queried together without defining any relationships in advance.
Build a dashboard relating sales calls from Salesforce to orders from Stripe. Or look at conversion rate from email opened in Mailchimp to a website visit in Segment.
All activities can be combined as soon as they're defined in the activity schema.