Grading Policies

Grading Weights

The final grade for this course depends upon the grades and scores earned on course components weighted as follows:

  • 5% : A number of in-class pop quizzes.
  • 5% : Individual paper reading assignment.
  • 5% : Team paper presentation.
  • 15% : Bi-daily reading assignments.
  • 15%: 5 individual homework.
  • 18%: 6 in-class labs.
  • 20%: 2 midterm exams.
  • 17%: Final Project - project proposal (10%). Milestone check (20%). Project demonstration and presentation (40%). Final project report and revised code (30%).

Grades will be computed by rounding numerical percentages to the nearest integer and applying the following table:

Letter Range Letter Range Letter Range
A 93-100 B- 80-82 D+ 67-69
A- 90-92 C+ 77-79 D 63-66
B+ 87-89 C 73-76 D- 60-62
B 83-86 C- 70-72 F 0-59

In-class labs

All labs are to be done individually, in class, with the assistance of the instructor. Each assignment will be given near the beginning of a class and must be uploaded to Canvas before the due time. Some of the labs allow for late submission till midnight with late penalty, but no submission is allowed beyond the lab day. To receive full credit, a programming assignment must have no code exceptions, but partial credit will be awarded based on the understanding of the material demonstrated by the student’s code.

Homework Assignments

All homework assignments have to be completed individually in this course. Homework assignments are meant to be more comprehensive tasks that may need additional reading and research. Homework assignments are to be completed in a week or two.

Late Policy

Late assignment will lose equation points per day, including weekends and holidays. For example, -1 after 1 day, -2 after 2 days, -4 after 3 days … until 8+ days students will receive 0 score. Under no circumstances should any student disclose their code, or copy the code written by other students or from an outside source. Both students will receive 0 score on the disclosed/copied assignment.

Paper Reading Assignment & Paper Presentation

Presenting data analysis results to public is one of the highly sought skills from data science employers.

Students are required to form a team of two and read at least one technical paper from the provided paper list (or other papers agreed by the instructor). Students need to complete 1) an individual paper reading assignment in the middle of the semester; and 2) team-present the paper to the instructor and their peers during the end of the semester. The grading of the paper presentation consists of 50% from peers and 50% from the instructor. The evaluation rubrics will be provided via Canvas.

Final Project

The final project best demonstrates students’ understanding of the lecture contents and their capability of solving real data science problems. Students are encouraged to form teams of no more than 2 people. A single solo project is allowed if class size is odd. Topics of the final project are left open for the teams to find, but the final choice has to be decided upon the instructor’s agreement. A good list of potential projects can be found at Kaggle and Public Datasets. Teams are responsible for finding the project of their interest and discussing the project feasibility face-to-face with the project manager, a.k.a., the instructor 1 week before the project proposal is due. Project formation is on a rolling basis, meaning formerly finalized projects cannot be taken again by another team. The grading of the final project consists of four major aspects:

  • 10%: the project proposal (including the conversation with the instructor a week before due);
  • 20%: mid-term project check (must have shown substantial exploratory data analysis);
  • 40%: the project demonstration; code and slides are due the night before the presentation day;
  • 30%: everything else including report; revised code and the report are due on the last day of the final exam week.

The requirements and grading rubric for each component of the project will be posted on Canvas. Under no circumstances should any student copy the code from an outside source. Source code similarity tools will be used to detect plagiarism.

Note:

There is a potential pitfall of using open datasets to introduce data science that some students want to leverage. I want to be crystal clear that the point of this project is to learn as much as possible. The point of this project is not to achieve the most intriguing results on the Kaggle leaderboard. For instance, if you copied some code you didn't understand and used it to get very compelling analytical results, I would consider this a very poor project. However, if you thoughtfully try various strategies, incorporate lots of interesting learning resources, and do a good job communicating with your peers regarding what worked and what didn't, I would consider that highly successful even if your results were not the best.

Previous
Next