Hackathon preparation activity
In this activity, you will begin working with a realistic scenario and dataset. The goal is not to design a solution yet, but to interrogate the data and consider a credible ML problem that could realistically be taken forward during the hackathon.
Download the Hackathon Resource Pack
The pack contains:
- Descriptions of the two scenarios.
- Datasets for each scenario.
- Presentation template to capture your thinking throughout the hackathon and use to present to stakeholders
You’ll use this pack during the preparation activity and continue using it throughout the hackathon sprints.
AIMLF Hackathon Resource Pack.zipOnce downloaded, open the files and keep them accessible for the rest of the hackathon.
Choose a scenario
- You will work in groups of up to 4 people
- Open the** scenario descriptions** from theHackathon Resource Pack.
- As a group, review the two scenarios and select one scenario to focus on.
- Make your decision based on the scenario only (not the dataset).
- Use the presentation template to capture your thoughts.
Interrogate the dataset and schema
From the Hackathon Resource Pack, open the dataset linked to your chosen scenario. Explore the data at a high level and record observations in your presentation template:
- Identify the grain: What does one row represent?
- Audit the schema: What does each column represent? Are fields numerical, categorical, or time-based?
- Look for signals: Which columns logically relate to the problem context?
- Check data quality: Do a quick visual scan for missing values, odd ranges, or inconsistencies.
- Consider preparation: What pre-processing or feature engineering might be needed?
The dataset is intentionally in good shape. Focus on understanding, assumptions, and design choices, not detailed data cleaning.
Before you move on
- Capture assumptions and open questions in your presentation template you will refine these during the sprint activities that follow.
- In the next stage of the hackathon, you'll build on this by developing and refining your AI/ML solution.