Graphbuilder: Scalable Graph Construction For Big Data Open Source Code Release
Graphs are powerful abstractions to discover hidden insights for applications from social media to business analytics, medicine and e-science. However, today it requires deep domain knowledge to build a graph which can be efficiently distributed across servers in the cloud. Developed by Intel Labs, GraphBuilder is the first scalable open source library to take large data sets and construct them into “Graphs,” web-like structures that outline relationships among data. By providing an extensible, general purpose graph-construction solution based on Hadoop, GraphBuilder aims to cut development time from months to days by eliminating the need to develop custom code. GraphBuilder completes an end-to-end machine-learning pipeline when combined with GraphLab, an open source Graph computing framework developed by CMU in association with our ISTC for Cloud Computing. GraphBuilder can be used to create Graphs from raw data, which then can be processed using GraphLab or similar tools.
Read Blog: GraphBuilder: Revealing Hidden Structure Within Big Data
Download: GraphBuilder is available at https://01.org/graphbuilder/ under the Apache 2 license.