During the first weekend of February 2015, the 15th edition of FOSDEM (Free and Open Source Software Developers' European Meeting) was organised at the Université Libre de Bruxelles.
With 5000+ attendees and more than 550 lectures, it is the place to learn about a lot of open source projects and their community. Since you can easily be overwhelmed by the hectic schedule, it makes sense to plan your visit in advance. Using one of the FOSDEM Apps can really help you with this.
It is impossible to join all of the interesting talks, so I will just mention some of the interesting talks we attended on the graph processing and the open source search track. Note that there are many more tracks and also lightning talks (e.g., the creator of AdBlock Plus talking about adblockers), so this is a highly skewed samples set of the talks given at FOSDEM.
Big Graph Analytics on Neo4j with Apache Spark
A Docker Image for Graph Analytics on Neo4j with Apache Spark GraphX
In this talk Kenny Bastani introduced a Docker container (Neo4j Mazerunner) that provides an easy way to do distributed graph processing using Apache Spark, GraphX, and a Neo4j graph database. He tested this docker image to do PageRank calculations. A more elaborate description of his work can be found on his blog: http://www.kennybastani.com
Recommendation Engines with Graph Databases
Building a high-performance recommendation engine using open-source software (Neo4j & GraphAware Framework)
The main committer of the GraphAware Framework, Michal Bachman, shared his experience building a number of high-performance production-ready recommendation engines using Neo4j and introduced the open-source GraphAware Recommendation Engine Library.
He highlighted why graphs are a suitable data model for building recommender systems, and also discussed some general business requirements that were taken into account when building the architecture of the Recommendation Engine Library. The source code can be found on GitHub: https://github.com/graphaware/neo4j-reco
Elasticsearch from the Bottom Up
During this talk about the architecture of Lucene and Elastic Search, Alex Brasetvik started with describing the inverted index, to ascend through abstraction layers to get an overview of how a distributed search cluster processes searches and changes. This deeper understanding is really helpful when debugging your Elastic Search environment.
ELK, making sense of your data (not just for logs!)
Elasticsearch, combined with Logstash and Kibana, is also known as the ELK stack, and can provide you with a great environment to analyse your data. During this talk Pere Urbon-Bayes showcased the stack by analysing the the eclipse project bugzilla repos.
All of the talks were recorded, and should be accessible on: http://video.fosdem.org/