# Kaggle
:::info
[TOC]
:::

<br/>

## Introduction
**Kaggle** is an online data science community and platform owned by Google.

It provides:
- 📊 **Datasets**: Public and private datasets.
- 🏆 **Competitions**: Machine learning challenges with leaderboards.
- 🛠 **Notebooks**: Free cloud-based Jupyter-like environment.
- 👥 **Community**: Forums and discussions.

> Perfect for beginners to practice ML and for experts to compete.

<br/>

## Create Account and API Key
1. Go to [Kaggle](https://www.kaggle.com/) and create an account.
2. After login → click profile icon → **Account**.
3. Scroll to **API** section → click **Create New API Token**.
- `kaggle.json`:

     {
       "username": "your_username",
       "key": "your_api_key"
     }

:::success
#### • Setup API Key on Windows
Place the `kaggle.json` file at ++"C:\Users\\\<Username\>\\.kaggle\kaggle.json"++

    """ Or set environment variables directly in Python """
    import os
    os.environ['KAGGLE_USERNAME'] = 'your_username'
    os.environ['KAGGLE_KEY'] = 'your_api_key'

:::

<br/>

## Install API & Download Datasets
Run in terminal (Anaconda Prompt / CMD):

pip install kaggle
kaggle --version

Common Download Commands:

# Search datasets
kaggle datasets list -s "titanic"

# Download a dataset
kaggle datasets download -d <dataset-owner>/<dataset-name>

# Example: download Titanic competition data
kaggle competitions download -c titanic

> Extract the `.zip` file after downloading.

<br/>

## Kaggle Resources

1. **Kaggle Competitions**
- Browse competitions: ++[Kaggle Competitions](https://www.kaggle.com/competitions)++
- Use Kaggle Notebooks or your local environment to build a model.
- Submit predictions as a `.csv` file → check your leaderboard ranking.

2. **Kaggle Notebooks**
Kaggle provides free cloud notebooks with:
- Jupyter-like interface
- Free GPU/TPU support
- Preinstalled ML libraries (TensorFlow, PyTorch, Scikit-learn, etc.)
- Direct access to Kaggle datasets
> Great for users without local GPU resources.

<br/>

- 💡 **Recommended Starter Competitions**
1. ++[Titanic: Machine Learning from Disaster](https://www.kaggle.com/c/titanic)++
- Predict passenger survival.
- Great for learning data cleaning, feature engineering, and classification.
2. ++[House Prices - Advanced Regression Techniques](https://www.kaggle.com/c/house-prices-advanced-regression-techniques)++
- Regression problem to predict house prices.
3. ++[Digit Recognizer](https://www.kaggle.com/c/digit-recognizer)++
- MNIST handwritten digits classification (intro to deep learning).

<br/>

## Summary
- Kaggle is the best platform to start learning data science.
- Always set up API (`kaggle.json`) for downloading datasets.
- Begin with the Titanic competition for hands-on practice.
- Learn from Kaggle Notebooks shared by the community.

<br/>

:::spoiler Relevant Resource
[Kaggle Datasets](https://www.kaggle.com/datasets)
[Kaggle Competitions](https://www.kaggle.com/competitions)
[Kaggle Notebooks](https://www.kaggle.com/code)
[Kaggle Learn Courses](https://www.kaggle.com/learn)
:::