We are curating the finalized list of projects for 2019, and this page will be updated accordingly. For now, we are listing the previous year's projects, some of which may be continued for further research opportunities with our 2019 cohort. For specifics of the 2018 cohort, please see the 2018 menu to the left.
Project 1: MOOC Visualization/Analytics
Mentors: Dr. Huzefa Rangwala, Jessica Lin and Jill Nelson
Develop a system that builds a visual dashboard for instructors and administrators. Data will be used from server logs of 5 MOOC classes obtained from Stanford Edx initiative and KDD Cup Dataset.
- Specific tasks will involve using a frequent mining algorithm to identify common sequential patterns of interaction with a MOOC.
Project 2: In-class Prediction Analysis
Mentors: Dr. Huzefa Rangwala and Aditya Johri
- model student learning within class for grade prediction using regression/deep learning on LMS data
- develop a visual analytics dashboard to help instructor provide patterns of success
- develop an early warning system for identifying students' survival in class
- Dataset to be used: Canvas Network Dataset
Project 3: Teaching Programming Strategies
Mentor: Dr. Thomas LaToza
- document programming strategies and see how crowd-programming can be usedful for learning.
- build a system for crowdsourcing strategy refinement
- Intelligent Autocomplete
- Investigate differing models of documents, evaluate performance
- build a tool that explains models back to the user
Project 4: Raising #STEM Awareness
Mentors: Dr. Aditya Johri and Hemant Purohit
- analyze twitter metadata about specific stem-related hashtags to determine who participates
- what kinds of messages are being shared
- what kind of networks form within and across hashtags
- build predictive models forecasting trends in STEM Awareness
Project 5: WELI: Assistive Wearables for Neurodiverse Students in Higher Education
Mentor: Dr. Vivian Motti
- track data on neurodiverse students receiving guidance during lectures via wearables.
- develop analytical models for identifying interventions.
- develop analytical models for determining success.
Project 6: Anomalies Detection
Mentor: Dr. Carlotta Domeniconi
- focus on the discovery of anomalies in graphs - how to properly construct graphs from data, the impact of different similarity measures among actors/entities, how to leverage attributes in graphs
- graphs vs hypegraphs: do hypergraphs help? how to represent hypergraphs?
- use insider trading data collected from publicly available data, and already processed as graphs