Breaking Down Our Billion User Reach with HyperLogLog
summary
Explains HyperLogLog, a probabilistic cardinality-estimation algorithm and data structure, and how it was used to compute and provide breakdowns of unique user counts across a network with around a billion users, addressing space and distribution challenges when measuring reach.
AI-written summary · about
related talks
12mScalable Uniques in Postgres - HyperLogLog
49mDistributed COUNT(DISTINCT) with HyperLogLog on PostgreSQL
56mSpark-driven Audience counting
01h 00mHyper Speed: When Big Data Blooms
41mProbabilistic Data Structures
48mHyper Speed! When Big Data Blooms
43mSegmenting 600 Million Users
16mLog Rehydration and Logs to Metrics