About This Hackathon

<p>Welcome to <strong>Week 5</strong> of the Weekly MachineHack Hackathon series! This week presents an exciting new data science challenge: creating a model to predict the volume of rented bikes at an hour level granularity for the given instances within the test data. This time series analysis problem opens up numerous possibilities for innovative and effective solutions as applicable within the mobility space and services. We eagerly anticipate your creative approaches and outcomes.</p><h3>Event Duration</h3><ul><li><strong>Start Date:</strong> 25 July 2024</li><li><strong>End Date:</strong> 15 August 2024</li></ul><h3>Challenge Details</h3><p>Your task is to develop a time series regression model capable of accurately predicting the volume of rented bikes at an hour level granularity for the given dates in test data. This problem requires participants to apply their knowledge of time series analysis and machine learning.</p><h3>Participation and Benefits</h3><ul><li><strong>Intermediate Level:</strong> The hackathon is geared towards participants with a basic understanding of machine learning and time series analysis.</li><li><strong>Community Engagement:</strong> Join our vibrant community on Telegram to discuss ideas, ask questions, and collaborate with fellow participants.</li><li><strong>Certificates:</strong> All participants will receive a certificate from MachineHack, and winners will be prominently featured on the leaderboard.</li><li><strong>Live Walkthrough Session:</strong> A live session will be held on <strong>31st July at 7 PM</strong> to guide participants through the challenge and offer valuable insights.</li></ul><h3>Submission and Evaluation</h3><ul><li><strong>Submission Format:</strong> Participants must submit their predictions in the specified format in the <strong>submission.csv</strong> file.</li><li><strong>Evaluation Metric:</strong> The models will be evaluated based on their <strong>Root Mean Squared Error (RMSE)</strong> in predicting the correct categories.</li><li><strong>Leaderboard:</strong> Stay updated on your progress and aim for the top spot on the leaderboard. Make sure to have the submission file corresponding to the best score as your latest submission before the hackathon expires.</li></ul><h3>Data Description</h3><p>The dataset for this hackathon includes:</p><ul><li><strong>training.csv:</strong> Contains the training data with the volume of rented bikes at an hosrcy level granularity.</li><li><strong>test.csv:</strong> Test data that participants will use to generate predictions.</li><li><strong>submission.csv:</strong> The format for submitting your predictions.</li></ul><h3>How to Crack This Challenge</h3><p>Here are some steps to get started:</p><p>1. <strong>Data Preprocessing and Feature Engineering:</strong> Handle missing values, create time-based features, and extract lag and rolling statistics.<br>2. <strong>Time Series Decomposition and Trend Analysis:</strong> Decompose into trend, seasonality, and residuals; remove trend and seasonal components.<br>3. <strong>Modeling Approaches:</strong> Use ARIMA, SARIMA, or ETS for statistical modeling, and Random Forest, XGBoost, LSTM, or GRU for machine learning and deep learning.<br>4. <strong>Validation and Hyperparameter Tuning:</strong> Employ Time Series Cross-Validation and optimize model parameters using Grid Search, Random Search, or Bayesian Optimization.</p><p>For our subscribers, a starter notebook will be available to help you kickstart your solution. This notebook provides a basic framework for data preprocessing and model building, which you can further enhance and customize.</p><h3>Getting Started</h3><ul><li><strong>Register Now:</strong> Ensure you are registered for the event to participate and receive updates.</li><li><strong>Download the Dataset:</strong> Access the dataset from the MachineHack platform to begin working on your solution.</li><li><strong>Join the Community:</strong> Engage with fellow participants and mentors through our Telegram group for support and collaboration.</li></ul><h3>Support and Resources</h3><p>For any questions or assistance, feel free to contact the support team at support@machinehack.com. Stay updated with the latest information and announcements by subscribing to our newsletter.</p><p>Happy Hacking and Growing! 🚀</p>

Welcome to Week 5 of the Weekly MachineHack Hackathon series! This week presents an exciting new data science challenge: creating a model to predict the volume of rented bikes at an hour level granularity for the given instances within the test data. This time series analysis problem opens up numerous possibilities for innovative and effective solutions as applicable within the mobility space and services. We eagerly anticipate your creative approaches and outcomes.

Event Duration

Start Date: 25 July 2024
End Date: 15 August 2024

Challenge Details

Your task is to develop a time series regression model capable of accurately predicting the volume of rented bikes at an hour level granularity for the given dates in test data. This problem requires participants to apply their knowledge of time series analysis and machine learning.

Participation and Benefits

Intermediate Level: The hackathon is geared towards participants with a basic understanding of machine learning and time series analysis.
Community Engagement: Join our vibrant community on Telegram to discuss ideas, ask questions, and collaborate with fellow participants.
Certificates: All participants will receive a certificate from MachineHack, and winners will be prominently featured on the leaderboard.
Live Walkthrough Session: A live session will be held on 31st July at 7 PM to guide participants through the challenge and offer valuable insights.

Submission and Evaluation

Submission Format: Participants must submit their predictions in the specified format in the submission.csv file.
Evaluation Metric: The models will be evaluated based on their Root Mean Squared Error (RMSE) in predicting the correct categories.
Leaderboard: Stay updated on your progress and aim for the top spot on the leaderboard. Make sure to have the submission file corresponding to the best score as your latest submission before the hackathon expires.

Data Description

The dataset for this hackathon includes:

training.csv: Contains the training data with the volume of rented bikes at an hosrcy level granularity.
test.csv: Test data that participants will use to generate predictions.
submission.csv: The format for submitting your predictions.

How to Crack This Challenge

Here are some steps to get started:

1. Data Preprocessing and Feature Engineering: Handle missing values, create time-based features, and extract lag and rolling statistics.
2. Time Series Decomposition and Trend Analysis: Decompose into trend, seasonality, and residuals; remove trend and seasonal components.
3. Modeling Approaches: Use ARIMA, SARIMA, or ETS for statistical modeling, and Random Forest, XGBoost, LSTM, or GRU for machine learning and deep learning.
4. Validation and Hyperparameter Tuning: Employ Time Series Cross-Validation and optimize model parameters using Grid Search, Random Search, or Bayesian Optimization.

For our subscribers, a starter notebook will be available to help you kickstart your solution. This notebook provides a basic framework for data preprocessing and model building, which you can further enhance and customize.

Getting Started

Register Now: Ensure you are registered for the event to participate and receive updates.
Download the Dataset: Access the dataset from the MachineHack platform to begin working on your solution.
Join the Community: Engage with fellow participants and mentors through our Telegram group for support and collaboration.

Support and Resources

For any questions or assistance, feel free to contact the support team at support@machinehack.com. Stay updated with the latest information and announcements by subscribing to our newsletter.

Happy Hacking and Growing! 🚀

RentalBikesVolumePrediction