Sometimes Geo Search isn't Geo Search
This is a followup to a talk we gave about how to solving our caterer search problem in Elastic Search. What makes our geo search particularly interesting / challenging is that it needs to be based on driving distance. We'll go over our attempts to port our search engine to Elastic Search and how/why we ultimately ended up deciding to keep the geo search in Postgres while moving metadata search to ElasticSearch. Along the way we'll see some of issues we ran into trying to use Nested Documents, Parent Child, and the challenges of a fully denormalized solution.
Jeff Dwyer is a software engineer who lives at the intersection of marketing and product. Previously at PatientsLikeMe and HubSpot, Jeff is now at ezCater helping connect people who need catering with absolutely delicious tacos. His focus is on experimentation, growth, AB testing and analytics. He lives in Somerville, MA and in his spare time he starts side projects with .it TLDs like http://forcerank.it and http://ratelim.it Bradley Harris has been taking things apart and putting them back together since he was a toddler. He has been doing the same with software platforms at Boston area startups for over 10 years. After working at Choicestream, Sermo, MapMyFitness, and Twine Health he struck out on his own as an independent consultant with an enthusiasm for doing the dirty work. Outside of work he enjoys running and directing local road races in Somerville, MA.
How to make enterprise-wide search not suck with Elasticsearch Enterprise search across many different line-of-business systems is hard, and too often it's deployed poorly, with users complaining that "search sucks". It doesn't have to be that way. Elasticsearch provides a great core for enterprise-wide search, but there's many ingredients that need to be added to it. This talk covers practices and techniques taken from multiple projects that use Elasticsearch to index content from multiple enterprise systems and surface it for a wide population of users. We've learned a ton and want to share those learnings, including approaches that worked well, and some mistakes and lessons learned. We'll also drill down in some technical depth on connectors and security.
Mike Gregory is the Director of Systems Engineering with BA Insight, focused on helping customers apply innovative technology to be successful with search. Mike has advised hundreds of Fortune 1000 companies and government agencies around the world. He has expertise in a wide variety of application areas including Enterprise Search, Knowledge Management, Enterprise Learning, Content Management, CRM, Call Center Management, and Human Resources. Mike is also an avid Red Sox and Patriots fan.