CS 49000-DSC: Data Science Capstone

Semester:	Spring 2022
Time and place:	Monday, Wednesday and Friday, 11.30am-12.20pm, Lawson Building 1142
Instructor:	Jean Honorio (Please send an e-mail for appointments)
TAs:	Nikhil Goyal, email: goyal70 at purdue.edu Yonghan Jung, email: jung222 at purdue.edu Jiacheng Li, email: li2829 at purdue.edu Hasan Mahmood, email: mahmood6 at purdue.edu Tanmaya Udupa, email: tudupa at purdue.edu

The Capstone course aims at providing students with an opportunity to integrate their accumulated knowledge and technical and social skills in order to identify and solve a real-world data science problem, with a special emphasis on the application domain. A capstone project is sponsored by a corporate partner or by an academic research group.

The Capstone course serves as a final preparation for students entering into the profession. Students will conduct a team-based project through the entire data science pipeline, by following the six phases of the CRISP-DM (CRoss-Industry Standard Process for Data Mining) methodology. Students get experience in working as teams, participating in project planning, writing reports and giving presentations.

Learning Objectives

After successful completion of this course, a student will be able to:

Identify the business objectives in a complex real-world problem.
Develop appropriate data science goals and project plan.
Implement the appropriate data pre-processing and analysis models.
Evaluate the above models, and interpret results with respect to the business objectives.
Present the project outcomes through written reports and digital/oral presentations.

In this syllabus, the term "business" refers to the corporate partner or academic research group that sponsors the capstone project.

Prerequisites

CS 37300.

Textbooks

There is no official textbook for this class. We will follow the CRISP-DM (CRoss-Industry Standard Process for Data Mining) manual. In addition, any data science book will be fine, for instance:

Principles of Data Mining by David Hand, Heikki Mannila and Padhraic Smyth. (Free with Purdue ID)
A Course in Machine Learning by Hal Daumé III. (Free)
Pattern Recognition and Machine Learning by Christopher M. Bishop.
Machine Learning by Tom Mitchell.
Pattern Classification, 2nd Edition by Richard O. Duda, Peter E. Hart, David G. Stork.

Grading

Team participation:

Business understanding report: 10%
Data understanding report: 10%
Data preparation report: 10%
Modeling report: 20%
Evaluation report: 10%
Deployment report: 10%

Individual participation:

Quizzes/attendance: 5%
Presentation of team project: 20%
Critique to other teams during presentation: 5%

Projects

This semester, we thank the following academic research groups and corporate partners for sponsoring projects, and the DataMine^* and the Data Science Consulting Service^** for their invaluable help:

Team	TA	Meetings	Project
7	Jiacheng Li	TA: Thursday 7:00pm-8:00pm, online	Prof. Stanley Chan, Electrical and Computer Engineering: Chan's group aims at understanding the vulnerability of machine learning under adversarial image-based attacks such as color perturbations.
3	Yonghan Jung	TA: Tuesday 2:30pm-3:30pm, online	Prof. Yiheng Feng^**, Civil Engineering: Feng's group aims at using a freeway vehicle trajectory dataset to model the behavior of drivers. This would help autonomous vehicles to avoid collisions with surrounding vehicles and better plan their trajectories.
11	Hasan Mahmood	TA: Wednesday 4:30pm-5:30pm, online	Prof. Andrew Flachs, Anthropology: Flachs' group aims at analyzing hundreds of pages from interviews of farmers and farmer's market managers, to uncover adaptations to the pandemic and similarities/agreements.
1	Jiacheng Li	TA: Friday 1:00pm-2:00pm, online	Prof. Wen Jiang^**, Biological Sciences: Jiang's group aims to reconstruct the 3D structure of viruses from Cryo-EM: noisy 2D projection images from arbitrary, unknown viewpoints.
9	Jiacheng Li	TA: Monday 10:00am-11:00am, online	Prof. Guang Lin^**, Mechanical Engineering: Physics-based simulation is computationally expensive. Lin's group aims at using machine learning to produce similar but faster results than a simulator.
10	Yonghan Jung	TA: Monday 2:30pm-3:30pm, online	Prof. Sorin Matei, Communication: Matei's group aims at analyzing the bombing missions conducted by the allies in World War II, for tasks such as data reconstruction, prediction and intervention.
2	Hasan Mahmood	TA: Wednesday 2:00pm-3:00pm, online	Prof. Eric Waltenburg, Political Science: Waltenburg's group aims at analyzing opinions of the US Circuit Courts of Appeal to find out whether more complex (covering more issues) and polarized decisions relate to more diverse panels.
8	Tanmaya Udupa	TA: Friday 9:30am-10:30am, MRGN 112 Expert: Monday 9:30am-10:20am, online	Cat Digital^*: Caterpillar needs to categorize invoices from parts sales to customers, with improved results over existing natural language processing algorithms.
5	Tanmaya Udupa	TA: Tuesday 11:30am-12:30pm, MRGN 212 Expert: Thursday 12:30pm-1:20pm, online	Ford Motor Company^*: Ford aims to analyze voice of customer (social media, surveys, etc.) data by analyzing topic shifts over time to observe emerging trends.
12	Tanmaya Udupa	TA: Thursday 3:30pm-4:30pm, MRGN 206 Expert: Tuesday 3:30pm-4:20pm, online	Helmer Scientific^*: Helmer would like to better predict their ability to meet their forecast commitments and determine trends that could impact product mix in manufacturing.
6	Nikhil Goyal	TA: Friday 4:20pm-5:20pm, MRGN 212 Expert: Monday 3:30pm-4:20pm, online	Midcontinent Independent System Operator^*: MISO aims to use predictive analytics to allocate system resources (CPU, RAM, server, etc.) to calculate the optimal dispatch of generation to meet electricity needs.
4	Nikhil Goyal	TA: Friday 10:30am-11:30am, MRGN 112 Expert: Monday 11:30am-12:20pm, online	Republic Airways^*: Republic Airways aims to find markers or deviations in normal performance from the Fault History DataBase, Quick Access Recorder, and other telemetry, to predict future faults.

Late policy

Assignments that are submitted after the specified deadline will incur a 10% reduction in score per day late.

Academic Honesty

Please read the departmental academic integrity policy here. This will be followed unless we provide written documentation of exceptions. We encourage you to interact amongst yourselves: you may discuss and obtain help with basic concepts covered in lectures and homework specification (but not solution). However, unless otherwise noted, work turned in should reflect your own efforts and knowledge. Sharing or copying solutions is unacceptable and could result in failure. You are expected to take reasonable precautions to prevent others from using your work.

Additional course policies

Please read the general course policies here.

Schedule

Date	Topic (Tentative)	Notes
Mon, Jan 10	Course introduction CRISP-DM (CRoss-Industry Standard Process for Data Mining) methodology
Wed, Jan 12	Phase 1: Business understanding Case Study Report 1: Business understanding (password-protected)
Fri, Jan 14	lecture continues
Mon, Jan 17	MARTIN LUTHER KING JR. DAY
Wed, Jan 19	Phase 2: Data understanding Case Study Report 2: Data understanding (password-protected) (attendance by iClicker)	Business understanding report, due on Wed, Feb 2, 11.59pm EST (See Brightspace for directions)
Fri, Jan 21	lecture continues (attendance by iClicker)
Mon, Jan 24	Phase 3: Data preparation Case Study Report 3: Data preparation (password-protected) (attendance by iClicker)
Wed, Jan 26	lecture continues (attendance by iClicker)
Fri, Jan 28	—
Mon, Jan 31	Cross-validation (attendance by iClicker)
Wed, Feb 2	lecture continues Model selection (attendance by iClicker)	Business understanding report due ~~Data understanding report, due on Wed, Feb 16, 11.59pm EST~~ Data understanding report, due on Fri, Feb 18, 11.59pm EST (See Brightspace for directions)
Fri, Feb 4	lecture continues (attendance by Zoom)	online lecture, see Piazza for details
Mon, Feb 7	Student presentations of Business understanding report (attendance by iClicker)	see Piazza for details
Wed, Feb 9	presentations continue (attendance by iClicker)
Fri, Feb 11	presentations continue (attendance by iClicker)
Mon, Feb 14	presentations continue (attendance by iClicker)
Wed, Feb 16	presentations continue (only teams 5, 8 and 12)
Fri, Feb 18	Phase 4: Modeling Case Study Report 4: Modeling (password-protected) Case Study Code and Data (password-protected) (attendance by iClicker)	Data understanding report due Data preparation report, due on Fri, Mar 4, 11.59pm EST
Mon, Feb 21	lecture continues (attendance by iClicker)
Wed, Feb 23	Student presentations of Data understanding report (attendance by iClicker)
Fri, Feb 25	presentations continue (attendance by iClicker)
Mon, Feb 28	presentations continue (attendance by iClicker)
Wed, Mar 2	presentations continue (attendance by iClicker)
Fri, Mar 4	presentations continue (only teams 5 and 12)	Data preparation report due ~~Modeling report, due on Fri, Apr 1, 11.59pm EST~~ Modeling report, due on Wed, Apr 6, 11.59pm EST
Mon, Mar 7	presentations continue (only teams 4 and 8)
Wed, Mar 9	Student presentations of Data preparation report (attendance by iClicker)
Fri, Mar 11	presentations continue (attendance by iClicker)
Mon, Mar 14	SPRING VACATION
Wed, Mar 16	SPRING VACATION
Fri, Mar 18	SPRING VACATION
Mon, Mar 21	presentations continue (attendance by iClicker)
Wed, Mar 23	presentations continue (only teams 5 and 12)
Fri, Mar 25	presentations continue (attendance by iClicker)
Mon, Mar 28	presentations continue (only teams 4 and 6)
Wed, Mar 30	Office hours through Zoom (Team 1, 7 and 11 attended)
Fri, Apr 1	Office hours through Zoom (Team 5, 12 and 1 attended)
Mon, Apr 4	Phase 5: Evaluation Case Study Report 5: Evaluation (password-protected) (attendance by iClicker)
Wed, Apr 6	Phase 6: Deployment Case Study Report 6: Deployment (password-protected) (attendance by iClicker)	Modeling report due Evaluation report, due on Fri, Apr 15, 11.59pm EST
Fri, Apr 8	Student presentations of Modeling report (attendance by iClicker)
Mon, Apr 11	presentations continue (attendance by iClicker)
Wed, Apr 13	presentations continue (attendance by iClicker)
Fri, Apr 15	presentations continue (attendance by iClicker)	Evaluation report due
Mon, Apr 18	presentations continue (attendance by iClicker)
Wed, Apr 20	presentations continue (attendance by iClicker)	Presentation slides, due on Wed, Apr 27, 11.59pm EST
Fri, Apr 22	presentations continue (attendance by iClicker)
Mon, Apr 25	presentations continue (attendance by iClicker)
Wed, Apr 27	presentations continue (only teams 5 and 12)	Presentation slides due
Fri, Apr 29	presentations continue (only teams 6 and 8)	Additional reading materials: [1] and [2] (not mandatory to be read)