Predict Bike Sharing with AutoGluon in AWS SageMaker Studio

This repository contains a Jupyter Notebook (.ipynb file) that demonstrates how to use AutoGluon to predict bike sharing demand in AWS SageMaker Studio. AutoGluon is a powerful AutoML toolkit that automates the process of training and tuning machine learning models.

Table of Contents

• Prerequisites
• links
• Installation
• Dataset
• Usage
• AutoGluon Workflow
• Evaluation
• Cleanup
• References

Prerequisites

Before you begin, ensure you have the following:

• An AWS account with access to SageMaker Studio.
• SageMaker Studio set up in your AWS environment.
• Basic knowledge of Jupyter Notebooks and Python.

Links

Open in Google Colab

Installation

1. Clone the repository:

   git clone https://github.com/jahnvisahni31/predict_bike_sharing_with_autogluon.git
   cd predict_bike_sharing_with_autogluon

2. Open SageMaker Studio:

   Launch SageMaker Studio from the AWS Management Console.

3. Upload the Notebook:

   Upload the predict_bike_sharing_with_autogluon.ipynb file to your SageMaker Studio environment.

4. Install Required Libraries:

   Open a terminal in SageMaker Studio and run the following command to install AutoGluon:

   pip install autogluon

Dataset

The dataset used in this example is the Bike Sharing Demand dataset from Kaggle. You can download the dataset and upload it to your SageMaker Studio environment.

Usage

1. Open the Notebook:

   Open the predict_bike_sharing_with_autogluon.ipynb file in SageMaker Studio.

2. Follow the Steps:

   Follow the steps in the notebook to:

   • Load the dataset.
   • Preprocess the data.
   • Train the model using AutoGluon.
   • Evaluate the model's performance.

AutoGluon Workflow

The notebook demonstrates the following AutoGluon workflow:

1. Import Libraries:

   from autogluon.tabular import TabularPredictor

2. Load Dataset:

   import pandas as pd
   train_data = pd.read_csv('train.csv')
   test_data = pd.read_csv('test.csv')

3. Train Model:

   predictor = TabularPredictor(label='count').fit(train_data)

4. Evaluate Model:

   performance = predictor.evaluate(test_data)
   print(performance)

Evaluation

The notebook includes steps to evaluate the trained model on a test set, providing metrics such as RMSE (Root Mean Squared Error) to measure the model's performance.

Cleanup

After completing the notebook, remember to clean up any resources to avoid unnecessary charges:

• Delete any endpoints or instances created during the process.
• Remove datasets and notebooks from your SageMaker Studio environment if no longer needed.

References

• AutoGluon Documentation
• AWS SageMaker Studio Documentation
• Bike Sharing Demand Dataset on Kaggle

---

This README provides a high-level overview of using AutoGluon for bike sharing prediction in AWS SageMaker Studio. For detailed instructions and code, please refer to the included Jupyter Notebook.