Created a Google Code Project for FB Data Mining

I have created a project in Google code, project name is data-extraction-facebook, the project include extraction all code related to our Mining research will be checked in here. All the code is release under GNU Public license. Since this is a research project only team members have access to check-in, in that case most of the development will be done by myself and Mark Henning.

Facebook Application to Extract Data for Data Mining

After couple of stabs at facebook-java-api, I was able to build my frist facebook application which happened to be the first stage of developing a app to extract data from facebook for my Data Mining research. At Facebook has three types of RestClients that you can invoke to do your development. And in my case I use FacebookXMLRestClient.

At some point I will write a tutorial type blog post on building a facebook application, but for now I will have the screen shots. I have a feeling going forward I will run into an issue because, my intention is to populate the data model while I am extracting data. Since this is web approach I have to use Hibernate to get the database populate. For research project like this where I want to spend more time on developing our mining concepts rather than ORM technologies is not the best place to experiment my hibernate skills. So at somepoint I have to figure out away to make this as a desktop application and use regular JDBC connection to do data population.

As of now the application I wrote login and extract all the IDs of my Facebook friends, and the next step would be to iterate over each of my friend user id and get their friends.

Facebook Data Extraction

Here is the snapshot of the code, at very soon I will be hosting this project as a Google Code Project and will put the link out.

Looks like I will be more involve with the facebook-java-api team, so far I think we need more documentation on java api. Also I dont know why Facebook decided to stop supporting facebook java api.

Pulling my hair out on FB Data Extraction

I am trying it put together java base data extraction module for Facebook data. I am doing this to get large enough data set for my data mining project. I am thankful for all the developers who are trying to keep up the java-facebook-api upto date, but I had enormous amount of issues with facebook-java-api. Including my other research work this is the latest enemy of sleep. So, to give you an idea, for the last four days my total number of sleep hours are in single digit.

My Advisor is going to Kill me

I took some of my research work down here (Sri Lanka) thinking that I will get some work done while having fun.. but so far I am only having fun and I think I am going to continue doing that… If i am not wrong the time that I have put on my research work not more than days worth of work… what the heck… I am on vacation… Publication can wait… lets live the live for minute (very long minute…)

Finals are done, handed the first draft… I am heading home…

Its 4am in the morning and I just finished up the first draft of the publication to my research adviser along with some results of my program that i wrote to handle the nicknames issues of my research. I told my adviser that I will make an attempt to do work on my program while I am in Sri Lanka but I am afraid if I can really promise on it… hahah
I am so glad I am flying home (Sri Lanka) in about 8 hours… ooh god… feels so great… The breeze, the beach Sri Lanka I call it Paradise…

Objective::First phase of my research

Objectives
Main objectives of this research project are to implement Perl module to handle nick names as inputs and make suggestions for possible first names while attaching nick names with the matching last names to retrieve genealogical history of a given name, extend gen.pl along with additional modules to interface with http://www.ancestry.com to retrieve genealogical information related the names that are resolve by gel.pl module, add functionalities to handle combinations of errors such as misspelling, nicknames, mistakes during the census records and finally compare Soundex values of the return query and along with our implementation. The application will use gen.pl as its primary source and additional modules will use Perl as its primary language for implementation purpose.

gen.pl -> Pre-implementation by Dr. Bradley Rubin

Synopsis

I will be doing a presentation on

The Developmental Organization of Robot Behavior

Dr. Roderic A. Grupen
University of Massachusetts
Department of Computer Science

The book is not publish yet, but I will be doing a presenting the synopsis of the first chaper of this book. I am planning to show the video of the BigDog.

Presentation will be at HH302 2/08/2006 at 5pm……..All are welcome.