Arkajyoti Chakraborty, Director of Budget and Analytics, University of Virginia
Artificial Intelligence/data science has become widespread phenomenon in recent times. Every industry and sector tries to apply AI in its domain to solve problems. The range of AI applications vary from very fancy applications like ‘driverless car’ to daily operational task like ‘finding the next best customer’. Although higher education sector at large, lags the technological advancements like in many other sectors, there are instances where people are using data science applications to solve complex operational and research problems. Over time, these applications have been increasing in number.
Some of the areas where data science can be used are student admissions cycle, student lifecycle management, career placements, donor relations, financial operations and research/publications. Research itself is a broad area where there are several application of AI including deep learning (natural language processing and computer vision).
School admissions is a very critical area for any higher-education institute. On one hand, the institute has to make sure that the quality of students who get admitted is very high and on other hand, confirm that there should be enough good candidates to fill up the class size. The offer/acceptance ratio is important to hit a target where the institute has enough high-quality students to fill up the class size but do not have many more than the class size.
If too many applicants accept offers it will be a problem for Facilities Management to fit the class and then there can be a potential resource constraint issue. Financial aid/scholarship is offered to many applicants to entice them to accept the offer and usually, there is a fixed budget for financial aid. As seen above, the admission decision making has to include several moving parts and multiple constraints. This creates a great opportunity to use data science to solve many of these problems. Data science models can be used to predict who should be offered admissions, what is the chance of the admitted person to accept the offer and how much financial aid should be awarded for each potential offer to matriculate.
Universities get several data points for each applicant/ student from beginning of the application cycle, during several years of the program until career placement. These data comprise of academic scores, demographic, academic interaction, performance, placement and many more aspects. All these data can be used comprehensively to review, monitor and advise each student for better academic outcome during the course of the program and also in the future.
Data science models can be used to predict who should be offered admissions, what is the chance of the admitted person to accept the offer and how much financial aid should be awarded for each potential offer to matriculate
Beyond the student world, there are many other potential AI applicatons, which can improve the operations of the institution. One of the primary areas where most higher-education institutes have to focus is fund-raising. There are dedicated units within the school system who solely focus on fundraising. There are many historical data related to school alumni, their career profile, current job status, interests in alma mater, historical giving rate and historical endowment dollar amount. All these data are used to carve out a‘reach-out’ strategy to individual alum.AI can predict metrics like probability, amount, frequency and purpose of the endowment. Additionally, natural language processing can be used to automate the management of thousands of long fund documents and extract meaningful information like donor details, dollar amount, purpose of restriction and other important fund information.
Last but not the least; AI application can be used to streamline university financial operations. Machine Learning can be used to facilitate accounting in terms of coding, reconciliation and effective reporting.