Data Science Lab

This computation resource is a cluster of workstations that can work together as one big systems. Currently, the system can run large Hadoop and Spark jobs. There are also three GPU equipped nodes that are configured to run TensorFlow.

• System Description
• System Access
• How Do I
• Administration

HINT: To get back to this main page from any page in the wiki, click on the Data Science Lab in the upper left corner.

System News:

Feb-20-2019  Python Anaconda is now available, see the "How Do I" page for information on how to access it.
Nov-27-2018  Default Spark version is now 2.1.0, default Pyspark uses Python 3.6.3
Nov-07-2018: The Zeppelin Notebook is now available, see System Access above
Nov-01-2018: Python 3.6 updated on all systems with modules: numpy matplotlib TextBlob scipy 
             scikit-learn gensim pillow h5py xgboost happtbase mysqlclient happybase 
             (See "How Do I" for usage information) 
Jul-26-2018  R Studio server is installed. Enter  "http://localhost:8787" in a browser to access.
May-02-2018  R Libraries: See /opt/share/doc/Installing-R-Libraries for how to install your own R libraries.
Apr-20-2018: Python HBase lib HappyBase installed. Tensorflow now running on Limulus8-TF and Limulus9-TF
Feb-21-2018: A current Wikipedia snapshot is in HDFS at /data/Wikipedia
Feb-19-2018: HDFS is now available on all limulus machines as /mnt/hdfs
             Annotated examples from the purple Hadoop book are in /opt/share/doc/Hadoop2_Quick_Start_V1
Feb-14-2018: The following Python 2.7 modules are installed: nltk, keras, numpy, pandas, matplotlib
             TextBlob, scipy, Tensorflow, scikit-learn, gensim, pillow, h5py 
             !!! Be sure to run "scl enable devtoolset-6 python27 bash" to use Python 2.7             
Feb-09-2018: All external ssh connections will close after 30 minutes of inactivity. Internal
             ssh connections (machine to machine) will close after 1 hour of inactivity.