Data Science

With the massive data amount that can be generated and collected nowadays – so called big data, it demands a highly advanced skills set to analyze the data and derive new solutions for a vast number of applications.

To become a data scientist means to be “better at statistics than any software engineer and better at software engineering than any statistician,”  according to Josh Wills.

What is Big Data & Data Science?

• Big data: The next frontier for innovation, competition, and productivity
  • McKinsey’s public report, published in May 2011 but its content is still relevant
  • A thorough report on big data & data science techniques, technologies, applications, and market forecast
• Compare data science with other related terms
  • Distinguishing Analytics, Business Intelligence, Data Science
  • Data Science vs. Statistics vs. Business Intelligence
• 8 steps to become a data scientist – source: blog.datacamp.com

Common Techniques and Technologies

Below is a short (uncompleted) list of common skills required for undertaking data science tasks:

• Statistics, Probabilities, Statistical Inference, Statistical Modeling, and Data Visualization
• Mathematics, Discrete Math, Calculus, Linear Algebra, Numerical Analysis, and Algorithm
• Machine Learning
  • The study of computer algorithms that improve automatically through experience, by Tom Mitchell
  • Mitchell’s fundamental book on Machine Learning, published in 1997
  • Andrew Ng’s online course, including short video lectures
• Data Mining
  • The study of automatic sophisticated process that discover data patterns for segmentation and prediction
  • Free tools like WEKA, RapidMiner, KNIME, NLTK, Orange, and Apache Mahout.
  • Popular commercial tools like Microsoft  Azure, IBM SPSS Modeler, Rattle, MATLAB, SAS Enterprise Miner…
• Could platform like AWS, Azure, Cloudera
• Hadoop, HBase, Hive, Pig, Spark… 
  • Framework for processing large data sets with a cluster of smaller machines
  • Be familiar with Linux/Unix environment and shell scripts
  • Online Hadoop tutorials using Hortonworks Sandbox
  • Brief tutorials for Hadoop, HDFS, and MapReduce for experienced programmers
  • Introduction to MapReduce
• Programming languages like R, Python, Java (incl. Data Structures)…
• Structure Query Language (in general); relational vs. non-relational database
  • SQL tutorials with web practices
• Other statistical softwares like SAS, SPSS, STATA, Tableau, and QlikView…
• And soft skills like Domain Knowledge, Communication, Cooperation, Management, Creativity, Curiosity, and Ethics

Other Resources for Data Science

• Recommended books available ONLINE through Mason
• An Introduction to Data Science by Jeffrey Stanton, Syracuse University
• Free online courses
  • Coursera: Statistics and Data Science
  • edX: Data Analysis & Statistics
  • Indiana University: Information Visualization
• Some academic programs:
  • List of Master’s programs in Data Science in the U.S.
  • GMU’s M.S. in Data Analytics Engineering
    • Also consider graduate courses in CSI, Statistics, Mathematics, and CS departments

Community Support

• The industry’s online resource for big data practitioners
• Data Miners’ Blogs
• IBM’s Big Data and Analytics Hub
• IBM’s Big Data University
• Kaggle – home for prediction competitions and tutorial scripts (mostly in Python or R)
• Sample Projects from “Data Science for Social Good”