Knowledge mining using Hadoop

I want to do a project Hadoop and map reduce and present it as my graduation project. To this, I've given some thought,searched over the internet and came up with the idea of implementing some basic knowledge mining algorithms say on a social websites like Facebook or may stckoverflow, Quora etc and draw some statistical graphs, comparisons frequency distributions and other sort of important values.For searching purpose would it be wise to use Apache Solr ?

I want know

1. If such thing is feasible using the above mentioned tools, if so how should I build up on this little idea?
2. Where can I learn about knowledge mining algorithms which are e开发者_如何学Goasy to implement using java and map reduce techniques?
3. In case this is a wrong idea please suggest what else can otherwise be done on using Hadoop and other related sub-projects?

Thank you

1. Your idea probably should be developed a bit further. E.g. try to think what is exactly what you are trying to achieve.
2. If you are looking for an easy to learn data mining tool take a look at Hive. More advanced algorithms you may find in Mahout - an open source implementation of many data mining algorithms.

If you are looking for examples or ideas take a look at this project which demonstrates use of Hadoop and Hive in tracking trends on one of the most populart social website, that is on Wikipedia: https://github.com/datawrangling/trendingtopics