More information for CMPSCI 691MA, Social Media Analysis and
Computational Social Science

Course Format and Requirements

This seminar meets once per week for 2 hours (or slightly less). There will be 1-2 readings per week, mostly of research papers. There will also be suggested readings that give more background.

All students must do readings and write essays that summarize and react to the readings. We will post them on a class blog and read each other’s reactions. Students must also sign up to do one or two presentations during the semester. Grading will be based on these, plus class participation.

The 1 credit version of the course has the above as requirements. The course is 3 credits if you also do a final project, which involves a substantial research investigation along the lines of what could eventually become a paper worth submitting to a conference or journal. This requirement includes a presentation and final report.

Who this course is for

This is a PhD-level seminar in the computer science department. Undergraduates and masters students require instructor’s permission to join. Students from other departments are welcome to join as well, but again will need instructor’s permission.

We won’t have formal prerequisites, but a background in machine learning and/or statistical methodology is strongly suggested. Experience in computation-heavy data analysis areas like natural language processing or network science may be helpful as well. Since CSS is an interdisciplinary area, it’s impossible for anyone to be fully prepared, so we’ll try to introduce topics as they come up. You should have had exposure to regression, classification, and clustering algorithms; and ideally, a few of the following concepts: bag-of-words representations of documents; graph/node/vertex/edge/link network terminologies; the difference between O(N) vs. O(N^2); what a confidence interval is; and why randomized experiments are useful for causal inference.

If you don’t have any ML background, note that during this spring semester the CS department is offering a zillion different machine learning courses for many possible levels of students — please consider taking one of CMPSCI 348, 589, 689, or 688 during the semester.

Beyond technical preparation, we’d like to have a group of students with demonstrated interest and passion in this area.

For enrollment, if SPIRE doesn’t let you enroll, fill out an override request at

Other possible topics (we only got to a subset of them)

Please check out these other cool seminars being offered at UMass this semester:

Also check out the CSSI speaker series.


Some of these have papers, some do not. If you google for the author and title of a talk, sometimes you can find a related paper.

Tutorials (usually slides, sometimes video)