Getting Started with Kaggle Beginner

This lesson walks you through creating your Kaggle account, setting up your profile, exploring the platform, running your first notebook, and using the Kaggle API for programmatic access.

Creating Your Account

1. Visit Kaggle

   Go to kaggle.com and click Register.

2. Sign Up

   Register with your Google account or email. Google sign-in is the fastest option.

3. Complete Profile

   Add your display name, bio, location, and occupation. A complete profile is required to reach "Contributor" rank.

4. Phone Verification

   Verify your phone number to unlock competition submissions and GPU access.

Profile Setup

Your Kaggle profile is your data science portfolio. Make it count:

• Bio: Describe your interests, skills, and goals in data science
• Organization: Add your company, university, or "Independent" if self-learning
• GitHub/LinkedIn: Link your profiles to showcase your broader work
• Profile picture: Add a professional photo to build credibility

Kaggle Notebooks (Free GPU)

Kaggle provides free cloud-based Jupyter notebooks with powerful hardware:

• GPU: NVIDIA P100 or T4 — 30 hours per week for free
• TPU: Google TPU v3-8 — 20 hours per week for free
• CPU: Unlimited (no quota limit)
• RAM: 16 GB (CPU) or 13 GB (GPU)
• Disk: ~73 GB of storage per session
• Session limit: 12 hours maximum per session

Kaggle API

The Kaggle CLI lets you interact with Kaggle from the command line:

# Install the Kaggle API
pip install kaggle

# Set up credentials (download kaggle.json from kaggle.com/settings)
# Place it in ~/.kaggle/kaggle.json (Linux/Mac) or
# C:\Users\<username>\.kaggle\kaggle.json (Windows)

# List competitions
kaggle competitions list

# Download competition data
kaggle competitions download -c titanic

# List datasets
kaggle datasets list -s "covid"

# Download a dataset
kaggle datasets download -d heptapod/titanic

# Submit to a competition
kaggle competitions submit -c titanic -f submission.csv -m "My first submission"

Your First Notebook

1. Navigate to Notebooks

   Click Code in the left sidebar, then New Notebook.

2. Choose Settings

   Select Python or R, GPU or CPU, and notebook or script mode.

3. Write Code

   Start with a simple data exploration using one of Kaggle's built-in datasets.

4. Run and Save

   Click Run All to execute. Your notebook is auto-saved to your account.

# First Kaggle Notebook: Explore the Titanic dataset
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns

# Kaggle datasets are available at /kaggle/input/
df = pd.read_csv('/kaggle/input/titanic/train.csv')
print(f"Shape: {df.shape}")
print(df.head())

# Quick visualization
fig, axes = plt.subplots(1, 2, figsize=(12, 5))
df['Survived'].value_counts().plot(kind='bar', ax=axes[0], title='Survival Count')
df['Pclass'].value_counts().sort_index().plot(kind='bar', ax=axes[1], title='Passengers by Class')
plt.tight_layout()
plt.show()

Exploring Datasets

Browse Kaggle's massive dataset collection:

• Visit kaggle.com/datasets
• Use filters: file type, size, license, tags, and usability rating
• Sort by: hotness, votes, most recent, or most downloaded
• Preview data directly in the browser before downloading

Joining Your First Competition

Start with a "Getting Started" competition like Titanic: Machine Learning from Disaster:

1. Go to kaggle.com/competitions
2. Click on "Titanic - Machine Learning from Disaster"
3. Click Join Competition and accept the rules
4. Download the data or open a notebook directly
5. Build your model and submit predictions