Data Science — Structured Learning Log

Who: Rushit Tholiya · 2nd year B.Tech (CS) at Nirma University, Ahmedabad

Why: I want to build end-to-end ML projects — from raw data to a working model — and move into a data or ML engineering role after graduation. This repo is my daily proof-of-work.

How: Working through the CampusX 100-Days of ML course. For each lecture, I write my own notes before watching the solution, then compare.

---

How this repository reflects my learning

• I write notes in my own words before watching the solution video
• Every task has a comment block at the top: what I tried, where I got stuck, what I learned
• I track actual mistakes — not just the polished final answer
• Notebooks include my own additional experiments beyond the assigned tasks

---

Progress

Module	Lectures	Tasks Solved	Status
Python Basics	12 / 12	12 / 12	✅ Complete
NumPy	3 / 3	3 / 3	✅ Complete
Pandas	3 / 7	3 / 7	✅ Complete
matplotlib	2 / 2	2 / 2	✅ Complete
seaborn	2 / 2	2 / 2	✅ Complete
Data Analysis Process	0 / 3	0 / 3	⏳ Upcoming
SQL	9 / 9	9 / 9	✅ Complete
Statistics	1 / 14	1 / 14	⏳ In progress
Machine Learning	70 / 100+	—	⏳ In progress

Last updated: July 2026

---

Repo structure

/
├── 01-python-basics/
│   ├── notes/       
│   └── tasks/        
├── 02-numpy/
├── 03-pandas/
├── 04-matplotlib/
├── 05-seaborn/
├── 06-data-analysis-process/
├── 07-SQL/
├── 08-statistics/
├── 09-machine-learning/
└── datasets/

---

A few things I found genuinely interesting so far

• Python's *args and **kwargs — I wrote 6 small functions to test every edge case. Most useful when you genuinely don't know how many arguments a function will receive at call time.
• NumPy broadcasting confused me for 3 days. I finally understood it by drawing out the shape transformations by hand — see 02-numpy/notes/S14_Advanced_Numpy.ipynb

There's a pattern in all three libraries — I know they exist, I use them constantly, and I still blank on the exact syntax under pressure. So instead of googling the same thing for the 50th time, I built a searchable quick-reference site for each:

#	Library	What's inside	Live Preview	Source
1	🐍 Python	All 69 built-ins, core methods, runnable examples	🔗 Open	.html
2	🔢 NumPy	Array creation, indexing, broadcasting, most-used functions	🔗 Open	.html
3	🐼 Pandas	DataFrame/Series methods, groupby, merging, string & datetime accessors, file I/O	🔗 Open	.html
4	📊 Matplotlib & Seaborn	Most-used plot types, one-liners for quick recall	🔗 Open	.html

---

🎛️ Interactive ML Demos

Alongside notes and notebooks, I'm building small interactive apps to actually feel how ML concepts behave under different settings, instead of just reading about them:

• Linear Regression — Hyperparameter Tuning — a Streamlit app to tune hyperparameters (learning rate, iterations, etc.) and watch their effect on the model live → 🔗 Open the demo

• Decision Tree Classifier — Hyperparameter Tuning — a Streamlit app to tune hyperparameters (criterion, max depth, min samples split, etc.) and visualize the decision boundary live → 🔗 Open the demo

• Decision Tree Regressor — Hyperparameter Tuning — a Streamlit app to tune hyperparameters (criterion, max depth, min samples split, etc.) and watch the regression fit change live → 🔗 Open the demo

• Bagging Classifier — Base Estimator Comparison — a Streamlit app to tune bagging hyperparameters (n_estimators, max samples, bootstrap, etc.) and compare the decision boundary against a single base estimator → 🔗 Open the demo

• Bagging Regressor — Base Estimator Comparison — a Streamlit app to tune bagging hyperparameters (n_estimators, max samples, bootstrap, etc.) and compare the regression fit against a single base estimator → 🔗 Open the demo

• Random Forest Classifier — Hyperparameter Tuning — a Streamlit app to tune random forest hyperparameters (n_estimators, max features, bootstrap, max samples, etc.) and visualize the decision boundary live → 🔗 Open the demo

---

Environment

Python 3.11  ·  Jupyter Notebook / JupyterLab / Google Colab
NumPy  ·  Pandas  ·  Matplotlib  ·  Seaborn  ·  Scikit-learn

Clone and run locally:

1. Create a new folder on your system and open it in the terminal.
2. Clone the repository:

git clone https://github.com/Rushit004/data-science-journey-campusx.git

3. Move into the project folder:

cd data-science-journey-campusx

4. Install the required libraries:

pip install numpy pandas matplotlib seaborn scikit-learn jupyterlab

5. Launch Jupyter Lab:

jupyter lab

After running the above commands, Jupyter Lab will open in your browser, where you can explore all the notebooks and work.

About me

Rushit Tholiya 🔗 LinkedIn 🔗 GitHub profile

Live repo — updated as I progress. Feedback and suggestions welcome via Issues.