Projects
Projects built while learning
Case studies and coding experiments from my B.Tech in Data Science. Each one records what I attempted, what broke, and what I learned.
Case studies
Practical solutions, documented honestly
Problem definition
Defining the question first
Each case study starts with the problem as I understood it, the dataset I could access, and the constraints I was working under. Writing this down early helps me notice when the question has drifted.
Model training
The hurdles along the way
Training runs rarely go as planned. I record the preprocessing choices, the models I tried, and the errors that slowed me down, so the reasoning behind each decision stays visible.
Lessons learned
What I would change next time
The most useful part of a project is the review afterwards. I note what worked, what I would approach differently, and where the results should not be over-trusted.
Code experiments
Smaller scripts and explorations
Short scripts, exploratory data analysis notebooks, and benchmark experiments. Each is a learning exercise rather than a finished product.
Scripts
Small Python scripts
Focused scripts that test one idea at a time, such as cleaning a messy file or automating a repetitive analysis step.
Notebooks
Exploratory data analysis
Notebooks where I look at a dataset before modelling it, noting distributions, missing values, and the questions the data raises.
Benchmarks
Comparison experiments
Side-by-side runs of simple approaches on the same data. The aim is to understand trade-offs, not to claim a winning method.