The online course on Big Data Analysis from BerkeleyX on the edX platform started a re-run this week with a new focus. It now teaches students to program using Spark's Machine Learning pipelines and DataFrames.
CS110x: Big Data Analysis with Apache Spark is a four week course at intermediate level that opened on August 15 2016 and runs until September 12 is the successor to CS100.1x: Introduction to Big Data with Apache Spark and has the same overall goal of enabling students to learn how to use Apache Spark to perform data analysis. However, whereas the previous incarnation focused only on Spark programming using lower-level Spark abstraction and programming paradigm of Resilient Distributed Datasets the new version shows how to use Apache Spark Machine Learning libraries to analyze Big Data using DataFrames, Spark SQL, and Resilient Distributed Datasets. This will make of interest to students who have taken CS100.1x but are unfamiliar with Spark Machine Learning pipelines as well as the new cohort of students coming to the course for the first time.
The course is taught by Anthony D Joseph who is both Professor in Electrical Engineering and Computer Science and Technical Adviser at Databricks. The previous version of the course received positive ratings (average 4.2 out of 5 stars) and the consensus was that the weekly labs were the core of the course. The course assignments for this version include Prediction using Machine Learning algorithms, Collaborative Filtering, and Textual Entity Recognition exercises that teach students how to manipulate datasets using parallel processing with PySpark, Spark SQL, and Spark Machine Learning Pipelines. The lab exercises account for 84% of the grade, the other 16% coming from multiple choice quizzes and all assignments are due by September 12th, 2016.
The syllabus of the course is as follows:
Week 1: Big Data and Data Science
- Introduction to Big Data and Data Science - examples of how data science can leverage big data, and learn about the risks of performing data science without statistics
- Performing Data Science and Preparing Data - explore data science definitions and topics, and the process of acquiring and preparing data, understand the statistics of Exploratory Data Analysis
- Machine Learning - learn about Spark's machine learning libraries, ML and mllib
- Lab 1: Power Plant Machine Learning Pipeline data exploration and visualization, learn about Spark's Machine Learning Pipeline, and apply and evaluate several Machine Learning algorithms to answer a business question
Week 2: Performing Data Science
- Data Science Roles
- Data Quality
- Data Cleaning
- Statistical Inference - learn about estimation, bias, variability, data distributions and the Central Limit Theorem
- Lab 2: Collaborative Filtering on a Movie Dataset
Week 3:Apache Spark's Resilient Distributed Datasets
- Spark Low-Level Primitives - learn about Spark's Resilient Distributed Datasets, transformations, and actions, and Spark's shared variables
- File Performance - understand the considerations for the performance of file read and write actions
- Lab 3: Text Analysis and Entity Resolution - perform text analysis and entity resolution on Google and Amazon product listings using Spark
- Statistics - learn about relations, associations, trends, patterns, correlation, and regression
Although CS110x can be taken on its own, it is the second part of the three course X series. The introductory 2-week course, CS is currently underway but there is still time to join in this presentation with the advantage of becoming familiar with the PySpark environment and covering the basics.
blog comments powered by Disqus