CrUX Cookbook

Recipes for extracting insights from the Chrome User Experience Report (CrUX).

Guided Tour of CrUX

1. Counting the number of distinct origins
2. Searching for origins by their domain
3. Generating First Contentful Paint (FCP) histogram data for an origin
   • visualizing the histogram
4. Generating FCP histograms for each device
   • visualizing the histograms
5. Measuring proportion of device usage
   • Measuring proportion of connection usage
   • Measuring proportion of device and connection usage
6. Generating FCP histogram with normalized bin widths
   • visualizing the histogram
7. Generating timeseries of competitors' fast FCP
   • visualizing the timeseries

Contributing

Pull requests welcome!

• Standard SQL preferred
• Match formatting of existing recipes ("Format Query" on BigQuery doesn't always look good)
• Name your recipe using descriptive keywords (eg median-fcp.sql)
• Use lots of comments whenever needed to help users understand what you're doing, why you're doing it that way, and any limitations of your approach
• Queries that consume 10+ GB must be noted in the comments
• Assign PR to @rviscomi for review