Using Python for Big Data Workloads (Part 1)

In Part 2, we will look at Python for Spark (PySpark), Machine Learning, and deep learning in depth. In this first part, we'll go over the basics, some examples, and some tutorials to get you started.

Get the latest Python for your environment — Linux, OSX, and even Windows are supported.   There's a debate whether to finally move to Python 3.x; try it and see if it works for all your tools. Since my Hadoop installation has Python 2.7, I am going to use that for my work.

Python is great. I can run it for Machine Learning, websites, from NiFi, deep learning, and stitching together a lot of jobs. Using Apache Zeppelin, I can run Python and PySpark without installing it and tons of modules on my developer workstation.

Python Resources

• Python Against Phoenix

• Python Against Hive (GitHub)

Here is a simple example:

from pyhive
import hive
conn = hive.connect('myhiveserverisawesome.tim.com').cursor()
conn.execute('SELECT * FROM amwatertweetshive LIMIT 10')
print conn.fetchone()

• Using HDFS

• Python With MQTT for IOT (Option 2)

• HBase 101

• Spark and Python

• Python Development With Zeppelin 

• Introduction to Machine Learning With Apache Spark and Zeppelin

• Getting Started With Apache Zeppelin

• Sentiment Analysis With Python

• Sentiment Analysis and CoreNLP With REST API

• Machine Learning With Apache Hive and Apache Hivemall

• Apache HiveMall With Spark, Hive, and Pig

• Sentiment Analysis and NLP With Python

• Python Deep Learning

• Example Zeppelin Notebooks

• Enabling Apache Zeppelin and Spark for Data Science in the Enterprise

• Introduction to Data Science

• Apache Spark Crash Course

• Hadoop Crashcourse

• Apache Spark Using Python Class