Outreach

Helping academic counselors counsel "at-risk" students

Comment

Inspiration

56% of students who start at 4-year college drop by year 6 Part-time students 45% more likely to drop out of college 30% freshmen drop out after their first year of college

These statistics lead us to believe that students are not getting enough help from counselors based on their various attributes. Academic counselors need to be aware of what is going on in their students lives so they can advice accordingly.

What it does

Uses starfish dataset to make a predictive model and make a guess as to what the "retention rate" for students which is the likelyhood that a student will stay in college based on various attributed. It then goes ahead and gives the academic counselor a "scale" of where the student lies on "in good standing" to "in severe risk". It then goes ahead to analyze the attributes for this student and predict what "problems" the student might be facing and give a suggestion as to how to address it.

How we built it

We built a model to predict the "retention rate" based on dataset using scikitlearn and linear regression. We built a webapp using Spring MVC and have a friendly UI where academic counselors can go and look up students and it would give all the information about the student including the probability that they will stay in school and the problems they might be facing.

Challenges we ran into

We initially attempted to use the Watson API, but it could not handle the dataset. The web interface for watson discovery kept loading when we loaded it with a JSON file. So instead we decided to build our own predictive model.

Accomplishments that we're proud of

We are proud that we utilized so many languages in one project and this is something none of us have done before. We used python to build the model and predict the retention percentages. We then moved the data over to Java and the spring framework. We used HTML and CSS to do the front end of the web app. So in the end we used Python, Java, HTML, CSS.

What we learned

While trying to research how to integrate an API with our app, we learned the post and get methods. In order to execute the posts and gets we also had to became familiar with postman.

What's next for Outreach

Trying to get the Watson discovery API to predict the percentages for us.

Share this project:

Updates