So, if you are so inclined to do the whole crowdsourcing thing, please go vote for my talk at SXSW 2012 – Apache Mahout: Bringing Intelligence to Your App and then maybe I will see you at SXSW in 2012.

Surprise and Coincidence – musings from the long tail: Real-time decision making using map-reduce

I’ll be at BarCampRDU tomorrow.  I proposed two sessions, one on Hadoop and Mahout and one on Lucene and Solr.  I don’t think I really want to do both, but I would like to do at least one, so we’ll see what other people are interested in.

If you’re around and you want to talk about any of these things, track me down.

Boy, this could really come in handy in Open Source, especially projects like Mahout, Nutch and distributed Solr.  I find my biggest personal challenge on Mahout is access to computing resources.  I personally don’t have the financial backing to buy much time on Amazon EC2.  I have been scraping by, here and there, but find myself constantly wanting access to more capabilities.

Sigh.  Maybe I should put more ads on this site and use the funds for buying EC2 time.

Congrats to the Hadoop team!  Score one for Open Source!

1. Ted Dunning is now a committer!  Welcome Ted!
2. I put up a patch for a map-reduce ready (MRR?) version of a Naive Bayes classifier.  Still needs some work, but feedback is appreciated.  It includes a sample of running it against the 20 Newsgroups data.
3. Seems like their is someone new providing comments, etc. every day.  The mailing list continues to grow.
4. We have slotted 3 Google Summer of Code participants and had many excellent proposals.  It was very difficult to decide.
5. The Taste project is now officially accepted into Mahout and will be running in distributed mode in no time, I’m sure.
6. Karl is working on a hierarchical clustering implementation, amongst other things.

We are also trying to obtain compute resources for committers to use such that we can do testing/benchmarking on a cluster as opposed to some hodge-podge of resources any individual has access to at any given point in time.  If you are in a position to donate cloud time, please drop me a line.

Threw my name in the ring for BarCamp RDU today.  Haven’t been to BarCamp before, but Erik Hatcher suggested I go and check it out.

Also put in a Proposed Session of “Apache Mahout and Hadoop – Having fun with Map Reduce and distributed computing”.  Figure we talk about the basics of M/R, Hadoop and Mahout programming, look at some Mahout examples, run some code, maybe even go nuts and try setting up a distributed job across laptops just for the fun of it.  Was also thinking it might be fun to talk about Lucene/Solr, but figured one session was enough, especially since I am a BarCamp virgin.

Additionally, there were a fair number of inquiries about Mahout at ApacheCon EU and I got to meet Karl Wettin and Isabel Drost there (two Mahout committers). I went well over 3 years in Lucene Java land before meeting any of the other committers on Lucene.

Next, we added Jeff Eastman as a committer. Jeff has jumped in head first and is already helping out a lot.

Sean Owen has donated the Taste collaborative filtering project. We also made Sean a committer, and he is already contributing in other areas. We are still waiting to clear the legal hurdles here, but I think you will see Taste in the Mahout code base within the month.

Finally, it’s not official yet, but keep your eyes on Mahout for what should be another significant announcement of an NLP/ML project joining Mahout. I know, I know, I’m such a tease…

Mahout’s newest committer, Jeff Eastman, has a new blog on Mahout and Hadoop…