Hadoop and Neo4j

Hadoop is being widely used for processing big data and Neo4j is a popular open-source graph database. When doing social network analysis on big data, a “natural” thought is to use them together. Unfortunately, Neo4j cannot work directly on HDFS or HBase. Is it good to use them together for social network analysis of big data? If yes, any pros/cons and how to do it efficiently? Or shall we try other options, such as Hadoop + Giraph, or Spark + GraphX? Please share your ideas, and all suggestions or experiences will be appreciated. Thanks.

Anyway, to know more about how Neo4j and Hadoop can work together, I came across two presentations below, which might be interested to those who are doing social network analysis of big data.

Serious network analysis using Hadoop and Neo4j
http://neo4j.com/news/serious-network-analysis-using-hadoop-and-neo4j/

I Mapreduced a Neo store: Creating large Neo4j Databases with Hadoop
http://2013.berlinbuzzwords.de/sessions/i-mapreduced-neo-store-creating-large-neo4j-databases-hadoop