All projects
Machine Learning2026Coursework

Student Performance Analysis

Data preprocessing & exploratory analysis

A fully documented preprocessing and EDA pipeline on the UCI Student Performance dataset, preparing data to predict students' final grades.

GitHub

The problem

Before any model can predict a student's final grade, the raw survey data — 395 students, 33 mixed-type attributes — has to be inspected, cleaned, encoded and understood.

The approach

  1. 1Built a reproducible loading step with three fallbacks (UCI client, direct download, local copy).
  2. 2Audited data quality, missing values and outliers, and explored the distribution of the target grade G3.
  3. 3Analysed grade progression (G1 → G2 → G3), categorical distributions and correlations with the target.
  4. 4Encoded and scaled features into a model-ready dataset, and presented the findings.

Outcome

  • Model-ready dataset exported for downstream prediction
  • Every step explained: what, why and what we observed

Built with

PythonPandasNumPyscikit-learnMatplotlibSeabornJupyter