"IDEAS is allowing the Graduate School to pursue important analytic questions that we would never be able to explore unassisted. My office provided data in its rawest form, and IDEAS has run with it in ways I couldn’t have imagined. It’s a real tour-de-force. As a strong believer in the importance of data for strategic decision making, I am extremely grateful to IDEAS for making this tremendous resource available to us."
Sarah-Jane Leslie, Dean of The Graduate School
- Employed machine learning to estimate the likelihood of post-graduate employment success of past doctoral applicants using data from applications prior to 2011
- Identified current applicants with highest chance of future employment success based on predicted probability calculations
UP NEXT: Determine if more advanced predictive analytics and new types of data can significantly improve the selection process
- Extracted data on doctoral and master’s-level graduate applications from various administrative systems and merged it with publicly available higher education data (e.g., IPEDS, U.S. News rankings)
- Used machine learning to highlight and analyze trends and key predictors of various outcomes
UP NEXT: Use predictive analytics to understand variations in graduate admission processes across the University.
- Collected information from the web on current employment for a select list of alumni
- Identified which alumni are currently employed as faculty at R1 or R2 institutions
- Matched this subset with records in faculty database developed by IDEAS to further validate employment information
UP NEXT: Develop additional machine learning techniques and identify other data sources to improve the graduate school’s ability to track alumni employment outcomes.
- Surveyed the landscape of available academic employment outcomes data and classified its use by institutions and researchers
- Developed a plan for systematically identifying, collecting, and formatting data on faculty and academic staff in universities and colleges nationwide
- Conducted a pilot to test our approach for data collection
UP NEXT: Compile data on faculty and academic staff at select institutions and match to existing student data sets to better understand academic career outcomes.
Analytics in Higher Ed
- Conduct exploratory analysis of publicly available data on funding for faculty
- Pilot a data mining project to compare internal data about specific corporate and foundation funding for Princeton faculty with external public information
UP NEXT: Evaluate the feasibility of expanding data mining techniques to collect external data about faculty funding