Objectives

Bit by ML is a machine learning competition hosted on Kaggle that bridges the gap between theoretical ML concepts and real-world application, focusing on "AI for Social Good." Participants work with epidemiological and climate datasets to build predictive models, emphasizing practical problem-solving, model evaluation, and reproducibility.

• Demystify Data Science: Demonstrate that ML is a tool for solving tangible human problems like disease prevention.
• Skill Development: Practice regression analysis, time-series forecasting, and data cleaning.
• Collaborative Learning: Encourage peer-to-peer learning through team participation and public kernel sharing on Kaggle.

Competition Flow

1. Launch: Problem statement release and Kaggle page goes live.
2. Learn: Introductory workshop explaining the dataset and baseline models.
3. Hack: One week coding period with live leaderboard tracking.
4. Submit: Final CSV submission and code repository upload.
5. Present: Top teams present their approach to judges.

Competition Structure

Rules

• Participation is open to FY/SY/TY/Final Year students — beginners to intermediate level.
• Register via Google Form; validated participants receive a Kaggle invite link.
• Submit a .csv file with non-negative integer predictions for total_cases.
• Performance is evaluated using Mean Absolute Error (MAE).
• Public leaderboard is calculated on 30% of test data; private leaderboard on the remaining 70%.
• Top 3 winners must present a code walkthrough — submissions matching existing GitHub repos will be disqualified.
• No external data usage beyond what is provided.

Calendar