# Crash Kaggle Competition Tutorial

> for [STAT3009: Recommender Systems](https://www.bendai.org/CUHK-STAT3009/)

Participating in a Kaggle competition can be a rewarding experience that helps you improve your data science skills. Here's a step-by-step tutorial to guide you through the process, along with examples.

### Create a Kaggle Account

1. **Sign Up**: Go to [Kaggle](https://www.kaggle.com/) and create an account if you don't already have one.
2. **Explore Competitions**: Navigate to the "Competitions" tab to see ongoing and upcoming competitions.

### Choose a Competition

1. **Select a Competition**: Pick a competition that interests you. For example, the "Titanic: Machine Learning from Disaster" competition.
2. **Read the Overview**: Understand the problem statement, evaluation metric, and data description.

---

> [!IMPORTANT]
> Here you have two methods to submit your solution:
> - Submit a Kaggle notebook ([see this section](#run-the-kaggle-notebook))
> - Run the code in Google Colab (or your local machine) and upload prediction to Kaggle ([see this section](#run-the-code-colab-or-locally))

---

## Run the Kaggle Notebook

1. **Create a New Notebook**:
- Navigate to the "Code" tab in the competition page.
- Click on "New Notebook" to create a Kaggle Jupyter Notebook.

2. **Load the Data**:

   # Load the training data directly from Kaggle datasets
   train_data = pd.read_csv('/kaggle/input/<your-competition-name>/train.csv')

3. **Data Exploration and Processing**: Use the notebook to explore and preprocess your data as needed.

4. **Make Predictions**: Run your prediction code directly in the notebook.

5. **Prepare Submission**:
In your notebook, create a DataFrame for your submission and save it as `submission.csv`.

   # Load the sample submission file to understand the required format
   sample_submission = pd.read_csv('/kaggle/input/<your-competition-name>/sample_submission.csv')
   submission = sample_submission.copy()
   
   # Create a DataFrame for your submission
   submission['<the-target-col-name>'] = <your-predictions>

   # Save to CSV
   submission.to_csv('submission.csv', index=False)

6. **Submit Your Notebook**:
- After running your notebook and generating predictions, click on the "`Save Version`" button at the top right of the notebook interface.
- Now your notebook is ready and saved, so you can go to the Notebook Viewer and jump to the "`Output`" section by selecting it from the right side of the page as shown in the figure below.
- Here you will find a "`Submit`" button in blue color, whenever you are ready to submit your notebook, click on that button as shown in the figure below.
- See this [link](https://www.kaggle.com/discussions/general/166755) for more details.

## Run the Code Colab or Locally

1. **Download the Data**: Navigate to the "Data" tab of the competition page and download the datasets (e.g., `train.csv`, `test.csv`).

2. **Set Up Your Environment**:
- For Google Colab, you can start a new notebook and upload your datasets or load them directly from Kaggle.
- If using a local machine, ensure you have Python and necessary libraries (like pandas, numpy, and matplotlib) installed.

3. **Load and Explore the Data**:

   import pandas as pd

   # Load the training data
   train_data = pd.read_csv('train.csv')
   print(train_data.head())

4. **Make Predictions**: After processing and analyzing the data, generate predictions based on your model for `test.csv`.

5. **Prepare Submission**:

   # Load the sample submission file to understand the required format
   sample_submission = pd.read_csv('sample_submission.csv')
   submission = sample_submission.copy()
   
   # Create a DataFrame for your submission
   submission['<the-target-col-name>'] = <your-predictions>

   # Save to CSV
   submission.to_csv('submission.csv', index=False)

6. **Submit to Kaggle**:
- Go to the "Submit Predictions" tab on the competition page.
- Upload your `submission.csv` file and click "Submit".

### Conclusion

Participating in Kaggle competitions involves data exploration and understanding how to submit your predictions effectively. You can either download datasets and submit a CSV file or use Kaggle's Jupyter Notebook environment to make and submit predictions directly. Happy Kaggle-ing!