Project Description

While completing a Masters degree in Information Management at The University of Washingtons iSchool, I worked with the schools Admissions Board to investigate the factors that led to students registering for the Master of Science in Information Management upon acceptance.

In undertaking this inquiry, I deployed admissions, demographics, and previous education datasets. I further analyzed the competitor landscape to give insight into the Information Schools at relevant institutions. My comprehensive report of competitors included the types of programs provided, the technical level of their curricula, the availability of internships and other experiential learning opportunities, and the job prospects and salaries for graduates.

Competitor analysis and student trends provided vital intel to inform strategic planning and program development efforts better, ensuring that the University of Washington’s iSchool remains competitive and attractive to prospective students.

During the initial ideation of this project, I came up with two ways of determining market competition. I decided to use a feature table and machine learning to determine factors that lead to students choosing other institutions. A feature table is commonly used in market research to compare and evaluate competing products or services based on predefined criteria. In the case of assessing market competition for the UW iSchool, a feature table could be used to compare the offerings of the iSchool with other similar programs in terms of specific features. Machine learning can also be used to determine factors that make students more likely to register for their first quarter of classes in the MSIM program.

Analysis

I deployed machine learning and statistical analysis to indentify the factors that made students more likely to register for their first quarter of class in the MSIM program. After testing different classification models, I found that the data best fit a simple Logistic Regression model, which I then used to determine relevant factors. For this work, the Logistic Regression allowed me to assess feature importance, as the values of the coefficients give an idea of feature importance.

The initial phase of this research project included using the feature table. The reason for choosing to focus on a feature table was to enable the data collection. I researched the top ten different institutions students decided to go to by using exit surveys from prospective students. Aspects of this research included price, deadlines, core classes, brand/ranking, online programs, number of specialties, and types of degrees. The determinations gave me great insight into how these other institutions are more competitive than the iSchool.

The feature table provided qualitative research and a numerical value to determine how competitive the iSchool is with other information school programs. It was vital for me to have a numerical value to help compare other institutions as this would provide a solid way for me to compare.

Next, I deployed machine learning and statistical analysis to identify the factors that made students more likely to register for their first quarter of class in the MSIM program. After testing different classification models, I found that the data best fit a simple Logistic Regression model, which I then used to determine relevant factors. For this work, the Logistic Regression allowed me to assess feature importance, as the values of the coefficients give an idea of feature importance.

While other models had better outcomes for accuracy, I decided to use a Logistic Regression model because of how easy it is for stakeholders to understand the model. The administration stakeholders for the iSchool didn’t have advanced knowledge of data science and mathematics. Therefore, it was easier to explain and visualize the model because of the linear relationship this model has. Choosing this method proved to have benefits in the discussions with stakeholders.

Outcomes

Using the feature table with the machine-learning model, I determined three aspects of the iSchool's competitiveness and ways to improve the curriculum, marketing, and scholarship opportunities. Providing the feature table enabled stakeholders to discuss what could be improved within the iSchool regarding our recommendations. These recommendations were rooted in the Logistic Regression feature importance. A quantitative analysis parallel to the feature table allowed me to make a data-driven decision. The following presentation was presented to MSIM iSchool Board Members, and all of the insights have been taken out and anonymized.

Next Steps

To understand these outcomes and ensure that they are rooted in truth. I want to analyze the exit survey data and change the survey questions to reflect the types of insights found in this research project. These surveys are optional and don’t provide much insight into the student body. Enabling the iSchool to use these surveys and quantitative analysis would ensure that the iSchool continues to be competitive and understand the pitfalls of the programs.

The subsequent iterations of this project will work to corroborate the above factors through exit surveys to prospective students who decline their admission offers. Given the data and resources to continue, this work would further analyze the competition and create an operational dashboard to report real-time data for wide-reaching visibility.