Apache: Big Data 2015

Many search applications simply suggest queries or data items with a prefix match on what the user is typing sorted by popularity. With the advent of the mobile era, local.ch becomes one of the most used applications for local search in Switzerland. Some characteristics of a mobile application e.g. (a) difficulty of typing on small screen, and (b) more available personalized data from users (e.g. location), generated challenges for us when implementing the autosuggest feature. We finally brought a better autosuggest experience to our users based on: their search history, popularity of internal and external data, user context awareness (language, location), spellchecker...
In this talk we will describe the architecture of our autosuggest feature, how did we use Solr as a search component and Elasticsearch as a data aggregation component to improve autosuggest in our mobile application.