CSTrackOptimizerReact

Problem

At Purdue University, there are a a lot of tracks that students can accomplish in the Computer Science Program. While many students only end up completing one track, others decide to complete 2 or more tracks. To do so involves a significant amount of thought into what courses should be taken to complete the selected tracks. Let's consider the following example: Machine Learning/Intelligence and Software. Even though either Compilers or Operating Systems can be used for Software, Compilers should be chosen because it can intersect with Machine Learning/Intelligence as an elective. This kind of thinking is not very straightforward and gets more abstract as more tracks are selected.

Solution

To rectify this problem and make the Computer Science Track proccess at Purdue easier, I developed an algorithm that can find the minimum number of courses required to complete any number of tracks. Also, it allows for some modularity by allowing the user to select courses that they wish to take or have already taken. This feature allows students who wish to pivot from one track to another an easier time with selecting the right courses given their already selected ones.

Course Priorities

The course selection algorithm takes the following into consideration when courses are being recommended.

1. Smallest number of courses
2. Course Availability (How often the course shows up in rotation)
3. Course number (300s have higher priority than 400s)

How it works

The courses are separated into 3 subcategories for each track

1. Static Requirements
2. Dynammic Requirements
3. Electives

Static requirements are automatically added to the course list because they have to be there no matter what. I allocate the dynammic requirements using Catesian Products adn find the tuple that has the smallest number of courses. In the event of a tie, the tuples that tie are all considered as possible solutions. Lastly, the electives are allocated using a decision tree based algorithm traversed a Breath First Search method. This will help in finding the smallest number of courses. Once the decision tree creates electives in such a way that the algorithm finds the smallest number of courses, the answer is found and the result is displayed.

How to use algorithm/program

The serverless_client is being hosted using Github pages and can be found using the following link: https://athata1.github.io/CSTrackOptimizerReact/

Future Improvements

1. Improve algorithm in both speed and accuracy
2. Add limited ability to remove courses
3. Reimplementation of this project in a dynammic framework such as React or Angular

What I learned

1. Translating my previous java implementation into javascript
2. How to cohesively use data structures and algorithms to solve an abstract problem
3. Utilizing classes and inheritance to better contain information as well as add general functionality to all programs