Big Data Analytics (EECS E6893) and Advanced Big Data Analytics (EECS E6895)





E6893 Student List and Group Assignments
Big Data Public Dataset Information


COURSE BENEFITS:

  • Students will gain knowledge on analyzing Big Data. It serves as an introductory course for graduate students who are expecting to face Big Data storage, processing, analysis, visualization, and application issues on both workplaces and research environments.

  • Gain knowledge on this fast-changing technological direction. Big Data Analytics is probably the fastest evolving issue in the IT world now. New tools and algorithms are being created and adopted swiftly. Get insight on what tools, algorithms, and platforms to use on which types of real world use cases.

  • Get hands-on experience on Analytics, Mobile, Social and Security issues on Big Data through homeworks and final project

  • Final Project Reports will be published as Proceedings and Final Project Software will become Open Source. (Sapphirine Big Data Analytics Open Source Applications last release: May 2017)


PROFESSOR CHING-YUNG LIN:


    Dr. Ching-Yung Lin is with Graphen, Inc. He was the IBM Chief Scientist, Graph Computing, and an IBM Distinguished Researcher. He led the Network Science and Machine Intelligence Department in IBM T. J. Watson Research Center. He is an Adjunct Professor in Columbia University since 2005, and was an Affiliate Professor in the University of Washington from 2003 to 2009 and an Adjunct Professor in New York University (NYU) in 2014.

Dr. Lin was elevated to IEEE Fellow in Nov 2011, the first IEEE Fellow in the area of Network Science. He is an author of 170+ publications and 26 awarded patents. In 2010, IBM Exploratory Research Career Review selected Dr. Lin as one of the five researchers "mostly likely to have the greatest scientific impact for IBM and the world.” His “Big Data Analytics” course in Columbia University is the Top 1 search result of Baidu search on Big Data Analyticss.

In 2012-2015, he led a team of ~40 researchers from Columbia University, CMU, Northeastern Univ., Northwestern Univ., UC Berkeley, Stanford Research Institute, Rutgers Univ., Univ. of Minnesota, and NMU in the largest US social media analysis project including 26 tasks from 2012 to 2015. In 2015, he was invited to be a panelist together with the White House Chief Data Scientist in the semi-annual conference of the American Medical Association. He was invited as a keynote speaker in 20+ conferences, including the Expo 2.0 in New York Javits Convention Center in 2009. He was among the earliest researchers driving Machine Learning in Computer Vision, initiated the first large scale video annotation project by 111 researchers in 23 worldwide institutes in 2003. His work won 7 best paper awards and was featured 4 times by the BusinessWeek magazine, including being the Top Story of the Week in May 2009, the Best Paper Awards on ACM CIKM 2012 and IEEE BigData 2013.

Dr. Lin's team is now focusing on building novel software platform to simulate functions of brains to build various AI solutions in all kinds of industrial sectors, especially initially focusing on the Finance Industry.

 


APPLICABLE DEGREE PROGRAMS:

  • Recommended for MS or Ph.D. students in Electrical Engineering, Computer Science or any discipline requires big data analytics.
  • Most courses 4000-level and above can be credited to all degree programs.  All courses are subject to advisor approval.

COURSE FEES:

  • None


EECS E6893: Big Data Analytics

Lecturer/Manager:

Ching-Yung Lin

 

Office Hours:

Thursday 9:30 - 10:00pm or by appointment

Office Location/Phone:

SIPA 417

Email Address:

c {dot} lin {at} columbia {dot} edu

 

Day & Time Class 
Meets on Campus:

Thursday 7:00pm - 9:30pm

Location:

International Affairs Building (SIPA building) 417

Credits for course:

3

Class Type:

Lecture


Prerequisites:

This will be a hands-on course. Students need to know at least one or more programming languages: C, C++, Java, Perl, Python, and/or Javascript to finish homeworks and final project.


Description:

With the advance of IT storage, pcoressing, computation, and sensing technologies, Big Data has become a novel norm of life. Only until recently, computers are able to capture and analysis all sorts of large-scale data from all kinds of fields -- people, behavior, information, devices, sensors, biological signals, finance, vehicles, astronology, neurology, etc. Almost all industries are bracing into the challenge of Big Data and want to dig out valuable information to get insight to solve their challenges.


This course shall provide the fundamental knowledge to equip students being able to handle those challenges. This discipline inherently invoves many fields. Because of its importance and broad impact, new software and hardware tools and algorithms are quickly emerging. A data scientist needs to keep up with this ever changing trends to be able to create a state-of-the-art solution for real-world challenges.


This Big Data Analytics course shall first introduce the overview applications, market trend, and the things to learn. Then, I will introduce the fundamental platforms, such as Hadoop, Spark, and other tools, e.g., Linked Big Data. Afterwards, the course will introduce several data storage methods and how to upload, distribute, and process them. This shall include HDFS, HBase, KV stores, document database, and graph database. The course will go on to introduce different ways of handling analytics algorithms on different platforms. Then, I will introduce visualization issues and mobile issues on Big Data Analytics. Students will then have fundamental knowledge on Big Data Analytics to handle various real-world challenges.


Afterwards, the course will zoom in to discuss large-scale machine learning methods that are foundations for artificial intelligence and cognitive networks. The course will discuss several methods to optimize the analytics based on different hardware platforms, such as Intel & Power chips, GPU, FPGA, etc. The lectures will conclude with introduction of the future challenges of Big Data, especially on the onging Linked Big Data issues which involves graphs, graphical models, spatio-temporal analysis, cognitive analytics, etc.


Students will choose the topics of their own for a final project. The application domain can be based on the students' own interest. This will be a good opportunity for students to apply what's learned in the class for their needs, either for the future work requirements or for the research problems at hand.


TAs (Graders):

Gongqian Li , Sun Mao , Sheallika Singh , Vibhuti Mahajan , Vishal Anand , Aman Shankar , Congying Qiu , Lingqing Xu , Mohneesh Patel , and Ludwig Zhao

Office Hours:

Signup Sheet (Course Students only)

Office Location/Phone:

CS TA room

 

Required Textbook(s):

None

Reference Textbook(s):

class notes, and reference books or papers

Homework(s):

Three assignments (HW#1 - HW#3) including programming and written reports.

Project(s):

Final project in which students conduct research and hands-on implementation for self-selected topic on Big Data Analytics. Team collaboration of up to 3 students is encouraged.

Paper(s):

Report for each homework and the final project result. oral presentation of the final project proposal, intermediate presentation and final presentation results required.

Midterm Exam:

None

Final Exam:

None

Grading:

Three homework assignments: 40%, Final Project (proposal, intermediate and final presentations, report, open source code, and presentation video): 60%

Hardware
requirements:

PC with Internet access. 

Software
requirements:

Students may use their preferred software (C, C++, Java, Python, Perl, and/or Javascript) on their computers to complete homework assignments. 

Homework
submission:



by submission through Columbia CourseWorks

Course Outline

Class Date

Class 
Number

Topics Covered

Assignment

Due

09/07/17

1

Introduction to Big Data Analytics

 

09/14/17

2

Big Data Platforms

HW #1 Data Store & Processing

09/21/17

3

Big Data Storage and Analytics

 

09/28/17

4

Big Data Analytics ML Algorithms

HW #2 Recommendation, Clustering, and Classification

HW #1

10/05/17

5

Big Data Analytics ML Algorithms II

10/12/17

6

Graph Analytics

HW #3 Graph Database and Machine Reasoning

HW#2

10/19/17

7

Linked Big Data -- Graph Computing

10/26/17

8

Machine Reasoning for Big Data

11/02/17

9

Final Project Proposal Presentations

HW#3 & Proposal Slides

11/09/17

10

Big Data Visualization

11/16/17

11

Final Project Intermediate Presentations

 

11/23/17

 

NO CLASS -- Thanksgiving Holiday

 

 

11/30/17

12

GPU and Big Data Analytics

 

 Presentation Slides

12/07/17

13

Cognitive Mobile Analytics

 

 

12/14/17

14

Big Data Analytics Workshop

 

Final Project Slides

 

 


EECS E6895: Advanced Big Data Analytics

Lecturer:

Ching-Yung Lin

 

Office Hours:

Thursday 9:30 - 10:00pm or by appointment

Office Location/Phone:

Mudd 535

Email Address:

c {dot} lin {at} columbia {dot} edu

 

Day & Time Class 
Meets on Campus:

Thursday 7:00pm - 9:30pm

Location:

IAB 417

Credits for course:

3

Class Type:

Lecture


Prerequisites:

This will be a hands-on course. Students need to know at least one or more programming languages: C++, Java, Perl, Python, and/or Javascript to finish task milestones and final project.


Description:

The Big Data Analytics area evolves in a speed that was seldom seen in the history. New Software and Hardware tools are emerging and disruptive. Furthermore, its boundary with Artificial Intelligence becomes blurring. We may no longer find a clear distinction on what is a Big Data Analytics problem and what is an AI problem.


In this Advanced Big Data Analytics course, we will devote to something new -- "How far could we achieve to build a brain that mimics human functions through the state-of-the-art computer science and electrical engineering technologies?" What we would like to discuss is not machines that play Games (Chess, Question & Answering quiz, or Go) or recognize voice and face, but how machines could possibly achieve what are unique to the human beings. Our brains can reason, can associate, and can memorize. We have feeling, emotions, ethics and morality, arts, and consciousness. We dream during the night.


In this course, students will conduct Research and Development on the tasks that shall collectively contribute to building intelligent machines that are like human, or more knowledgable than human, through analyzing Big Data. Most lectures will be divided by two parts. The first part will be the presentations by Prof. Lin or guest speakers to explain the potential Computer Science / Electrical Engineering technologies for building such machines. The second part will be students' presentations on their progress in the 4 areas: (1) Cognitive Robot, (2) Finance Robo-Advisor, (3) Knowledge Graphs, and (4) Software for Next-Generation Hardware Platforms.


TAs (Graders):

TBD

Office Hours:

TBD

Office Location/Phone:

CS TA room, Mudd building

 

Required Textbook(s):

None

Reference Textbook(s):

class notes, and reference books or papers

Homework(s):

None


Task:

Each student will need to sign in a task in one of these four areas: (1) Cognitive Robot, (2) Finance Robo-Advisor, (3) Knowledge Graphs, and (4) Software for Next-Generation Hardwares. Each area will have 15 tasks. Task lists will be announced in Lecture 1. Each task will have three milestones. Each milestone includes programming, presentation and a written report.


Project:

Final project in which students define a Big Data Analytics application and apply the software built in any combination of the 60 tasks in the class to accomplish the project. Team collaboration of 2 students per project is encouraged.


Paper(s):

Report for each milestone and the final project results. Oral presentations of each milestone and the final project. Source codes will be submitted to a course Github repository. Final project will also include a video presentation.


Midterm Exam:

None

Final Exam:

None

Grading:

Task Milestones: 60% (3 milestones, each milestone: 20%), Final Project (presentation and report): 30%, Class Participation: 10%

Hardware
requirements:

PC with Internet access. 

Software
requirements:

Depending on the task, students will need to use appropriate software (C++, Java, Python, Perl, and/or Javascript) on their computers to complete the task milestones and the final project. 

Task
submission:



by submission through a course website.

Class Date

Class 
Number

Lecture Topics

Student Presentations

01/19/17

1

Advanced Big Data Analytics to Build a Brain

01/26/17

2

Foundations toward Human-Like Artificial Intelligence

 

02/02/17

3

Next-Generation Hardwares (I) -- GPU

Cognitive Robot (I)

02/09/17

4

Social Machines

Finance Robo-Advisor (I)

02/16/17

5

New Chips for AI & Big Data

Knowledge Graphs (I)

02/23/17

6

Ethics Machines (I)

SW for Next-Generation Hardwares (I)

03/02/17

7

Ethics Machines (II)

Cognitive Robot (II)

03/09/17

8

Feeling Machines (I)

Finance Robo-Advisor (II)

03/23/17

9

Feeling Machines (II)

Knowledge Graphs (II)

03/30/17

10

Arts-Aware Machines

SW for Next-Generation Hardwares (II)

04/06/17

11

Reasoning Machines (I)

Cognitive Robot (III)

04/13/17

12

Reasoning Machines (II)

Finance Robo-Advisor (III)

04/20/17

13

Consciousness Machines

Knowledge Graphs (III)

04/27/17

14

Dreaming Machines

SW for Next-Generation Hardwares (III)

05/11/17

15

Final Project Presentations