PredictEd

PredictEd is an academic analytics platform that combines machine learning with large language models to predict student outcomes and generate actionable insights for educators.

The prediction engine is a Python-based ML model trained on historical academic data — attendance, grades, assignment scores — to forecast end-semester performance with reasonable accuracy. On top of the predictions, Gemini and Grok LLMs are used to generate natural-language summaries and recommendations per student, surfacing patterns that raw numbers don't communicate.

The React frontend presents dashboards for teachers to review cohort-level and individual-level analytics. MongoDB stores student records and model outputs. The system is designed to be explainable — every prediction is accompanied by the key contributing factors.