In one recent article, we talked about benefits of graph databases over traditional databases (article: Graph Databases). This blog focuses on graph analytic techniques and how to deliver more relevant data.
Most common graph analytic techniques:
- Centrality analysis: identifying core entities in a network
- Community detection: finding clusters or communities
- Path analysis: identifying all connections between a pair of entities
- Sub-graph isomorphism is finding pattern of relationships
- Hadoop and graph analytics complementing techniques allow independent work items to be parceled in clusters (Hadoop) plus evaluating complex networks of relationships that cannot be segregated
Next some examples how graph theory is successfully used to find more relevance in complex data.
-
Google Knowledge Graph
The Google Knowledge Graph is the knowledge base that augments Google’s search engine’s search results. It combines classic search features with finding information through semantic search gathered from a wide range of different sources.
The Knowledge Graph allows users to search for people, things and places. Search results for these kinds of queries will include a variety of information relevant to the query such as celebrities, sports teams, movies, objects, images works of art, landmarks, cities, buildings and more.
What is so very innovative of the Google Knowledge Graph is that it aims to completely understand the context of the search request. That means that the search engine in leveraging as much data as about the user (e.g. location, age, former search preferences), the user’s interests and relationships.
-
Facebook Graph Search
Facebook Graph Search is a semantic search engine, searching based on intended meaning, responding to users’ natural language queries including people, pages, places, check-ins of the user, friends, where friends have been tagged, objects with location information, content, posts and comments created by the user or friends etc.
It combines the big data and metadata that were collected from its over one billion users and additional external data into a search engine providing user-specific search results.Rather than returning results based on matching keywords, the search engine is matching phrases, as well as objects on the site based on the users’ and their friends’ profiles and their relationships.
Summary
Graph analytics provide an incredible tool to cope with the huge amounts of social media and sensor-based data and, in addition, allows to uncover insights about the relationships between any information in the system to get new, more actionable and relevant answers to traditional and completely new questions.
