Facebook social data modeling with Scala, HBase, and HPaste
Scala Shell
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
bin
project
src
.gitignore
README.md
blend-techtalk-may-2012.pdf

README.md

social-modeling-starter

Facebook social data modeling with Scala, HBase, and HPaste.

Includes accessing the Facebook Graph API in Scala via Dispatch and Lift JSON, and using HPaste to access HBase and define MapReduce jobs.

Slides at http://blendlabsinc.com/blend-labs-tech-talk-stanford-may-2012.pdf.

Setup

Visit https://developers.facebook.com/tools/explorer?method=GET&path=me . Click "Get Access Token". Select the following permissions:

  • User Data Permissions
    • user_groups
    • user_likes
  • Friends Data Permissions
    • friends_groups
    • friends_likes

Click "Get Access Token", then "Allow". Copy and paste the Access Token. In your console, run:

export FACEBOOK_ACCESS_TOKEN=<your access token>

Download Cloudera's HBase distribution (CDH4 Beta 2) at http://archive.cloudera.com/cdh4/cdh/4/hbase-0.92.1-cdh4.0.0b2.tar.gz.

Unarchive the file and run

bin/start-hbase.sh
sleep 5
bin/hbase shell
hbase> create 'person', 'info', 'like'

To load your Facebook data, run

./bin/load.sh

Two MapReduce calculations are included: "friend likes" and "top likes". To execute these, run

./bin/topLikes.sh
./bin/friendLikes.sh