SMC Grade Predictor

Application of Linear Regression for Academic Outcome Forecasting: SMC Grade Predictor

Objective and Goals

The purpose of this project is to develop a grade prediction model that estimates a Santa Monica College student’s expected grade in a given math course taught by a specific professor.
The goal is to help students make informed class-selection decisions by using both historical grade data and student academic history, while also assisting instructors and counselors in identifying performance trends.

Requirements

Data Sources

  • SMC official grade distribution reports (past 3 years)

  • Student-submitted course histories and grades

  • Letter-grade-only dataset (A–F)

Technologies & Frameworks

  • Python

  • Scikit-learn (Linear Regression)

  • Flask (backend)

  • HTML/CSS/JavaScript (frontend)

  • Matplotlib (visualizations)

  • GitHub + Render for deployment

Planning

Milestones and Timeline

  • Collect and label historical grade data (701 classes, 22,425 grades)

  • Convert raw grade distributions into numerical values

  • Engineer two core predictors: Instructor Difficulty Score (IDS) and Student Ability Score (SAS)

  • Train linear regression model using historical + synthetic data

  • Build web application UI and integrate prediction API

  • Deploy on Render with live updates via GitHub

  • Evaluate model reliability using correlation significance tests

Roles and Responsibilities

Data Team

  • Gather grade distributions and student-provided data

  • Convert, clean, and format datasets for regression

  • Compute IDS and SAS metrics

Model Development Team

  • Implement linear regression pipeline

  • Train and validate model coefficients

  • Perform correlation significance analysis

  • Optimize model for reliability and consistency

Frontend Team

  • Build user interface using HTML/CSS/JS

  • Enable structured input for multiple previous courses

  • Display predictions and probability-like outputs

  • Implement dynamic features (add/remove courses)

Backend Team

  • Integrate Flask with prediction model

  • Process student inputs and return grade forecasts

  • Render Matplotlib charts (grade distributions, professor difficulty)

  • Maintain deployment pipeline through GitHub → Render

Execution

Development Process

  • Each member contributes weekly updates via GitHub branches

  • Python used for data modeling; Flask connects the model to the web app

  • Frontend collects course/professor history and desired class prediction

  • Backend returns predicted GPA and probable letter outcome

  • Matplotlib graphs display useful visuals (e.g., baseline grade distributions)

Weekly Meetings

  • Online meetings (30–60 minutes)

  • Discuss dataset updates, model accuracy, and UI improvements

  • Coordinate tasks using GitHub commits and version control