Finding relevancy: Improving your application search bar!
summary
Builds a small social‑network search using Elasticsearch + Kibana and a GitHub demo, starting from a single-field full‑text query then iteratively adding real features: terms lookup to boost first‑degree contacts, prefix search‑as‑you‑type on a full name field, field boosts for first vs last name, and optional rank_feature or vector fields for profit or similarity. Shows index mapping, the exact queries, and runtime scoring effects, and emphasizes logging clicks and conversion metrics to quantify relevancy improvements.
AI-written summary · about
chapters (19)
- 0:00Introduction to search relevancy
- 1:28Why search relevancy matters
- 8:50Query intent and user expectations
- 15:14LinkedIn search case study
- 19:48Elasticsearch basics
- 23:23Demo setup and overview
- 24:27Exploring the demo application
- 26:15Building the first search query
- 27:40Scoring direct contacts higher
- 28:46Boosting first and last names
- 30:34Adding full name search
- 31:17Implementing prefix search
- 32:44Handling name order variations
- 33:04Configuring the search index
- 35:30Iterating on search queries
- 36:34Key takeaways and best practices
- 39:25Advanced Elasticsearch features
- 40:30Measuring search experience
- 41:54Q&A session
related talks
37mBuilding a Search Engine to Improve User Engagement
46mFull-Text Search Internals with Elasticsearch
54mAdvanced (elastic)search for your legacy application
41mBeyond the basics with Elasticsearch
24mSearch From the Ground Up
26mFrom __icontains to search
51mAdvanced search for your legacy application
45mAnything New in Search? What If I Find It?