Algorithms 2 Repo

This repository contains a module of Jakob's Scheduler that predicts course enrollment. It works by training several precictions models on historical enrollment data and past schedules to make accurate predictions about future enrollment.

Django Webserver

This repository contains a simple Django webserver that runs inside a Docker container.

Prerequisites

• Docker

Build the Docker Image

docker-compose build

Run the Docker Container

docker-compose up

Documentation

The server exposes a single endpoint: /predict. By default, the django server runs on port 8000, but for compatibility with the rest of the scheduler, the docker container exposes port 8001.

/predict

http://localhost:8001/predict

This endpoint is used to make predictions about future enrollment. It takes the following parameters:

• year: int, the year for which to make predictions
• term: string, the term for which to make predictions ('fall', 'spring', 'summer')
• courses: Course[], A list of courses that will be included in the results

Course has the following structure:

{
	"course": string,
	"name": string,
	"prerequisites": string[][],
	"corequisites": string[][],
	"terms_offered": string[]
}

The server will return a JSON object with the following structure:

{
	"estimates": [
		{
			"course": string,
			"estimate": int
		},
		...
	]
}

Data

The repository contains a directory data/ that contains two subdirectories:

• client_data/: contains local copies of data that is used for testing purposes
• model_data/: contains temporary data that is used by the models
• scripts/ : contains scripts that were used to transalte between data formats

Enrollment Predictions

The repository contains a directory enrollment_predictions/ that contains the majority of the code. It contains the following subdirectories:

• models/: contains the code for the machine learning models
• modules/: contains the code for other modules that do not use machine learning
• tests/: contains the unit tests for the code

Other files in this directory are used by Django to run the server.

Data Visualization

The repository also contains a simple data visualization tool that can be used to visualize certain aspects of the data and models. It can be found in the Data Visualization/ directory.

Dependancies

• Python >= 3.9

This module also requires certain python packages that can be installed with:

pip install -r "Data Visualization/requirements.txt"

Running the Data Visualization Tool

python "Data Visualization/visualization.py"