Hackathon Instructions

Theme: Health and wellbeing or Migraine prediction & insights

Welcome to the Hackathon! In this repository, you will find many resources you may use, and instructions how to use them to get started with your hackathon project.

📁 Repository Structure

Data (data/)

The heart of the project with several different data sources to choose from:

1. MBRAIN21 Dataset (data/mbrain21-csv/)

• Real research data about migraines from Kaggle
• Contains daily event dumps for 4 patients over 3 months (MBRAIN21-001 to MBRAIN21-004)
• Could be useful for building prediction models based on real anonymised user data
• More info and credits: https://www.kaggle.com/datasets/jonvdrdo/mbrain21

2. Relief aHead API (data/relief ahead/)

• Neurawave API used for tracking personal wellbeing. Only own users data is available.
• Build your own backend based on this specification, or use it as inspiration for your data model
• API url: https://api.reliefahead.com/test
• API docs
  • Description: api-docs/README.md
  • API structure specification: openapi.yaml
  • You can also run Swagger to see API structure. Instructions in reliefaheadAPI.md Swagger a reference/example only - there is no actual localhost backend bundled with Swagger
• Get a user (username+password) with random data from hackathon organisers (hack1, hack2, hack3, ...)
• You can also create your own empty test user using Relief aHead Android app

3. Open Seizure Detector (data/openseizuredetector/)

• Open source project for health data management
• Can be customized and further developed even after the hackathon
• GitHub: https://github.com/OpenSeizureDetector/webApi/tree/master/api
• Website: https://www.openseizuredetector.org.uk/
• Note: Repos contains code to collect data. Contact hachathon organizers about pre-created data.

4. Synthetic Data (data/synthetic/)

• Large amounts of synthetic (generated) data for testing
• Two sizes: 10,000 and 100,000 persons
• Contains:
  • person_data_*.csv - Personal information
  • health_data_*.csv - Health data with triggers and migraine events
  • weather_data.csv - Weather data
• Detailed documentation of all fields in synthetic.md
• Perfect for training machine learning models without privacy concerns
• Download: The CSV files are too large to include in this repository. Please download them manually:
  • synthetic_data_10_000.zip (108 MB)
  • synthetic_data_100_000.zip (1.09 GB)

5. Korean S1 Dataset (data/korean-s1-dataset/)

• Research data about trigger factors in episodic migraineurs using a smartphone headache diary application
• Published study with 62 participants over 4579 days
• More info in korean-s1-dataset.md
• The original article: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0149577

Example applications

1. Mobile app example (flutter_app/)

A ready-made Flutter app with Bootstrap-inspired design that you can use as a foundation:

• Support for web, iOS and Android
• Run with flutter run (see README.md for details)

2. Fetching weather (api_examples/)

• weather_api_example.ipynb - Jupyter notebook with examples of how to fetch weather data from SMHI
• Can be used to connect weather data with migraine triggers

3. API Backend examples (backend/)

• A simple Flask API example that you can use as a starting point for building your own backend
• Ready to run with just pip install -r requirements.txt and python main.py
• Includes a basic "Hello World" endpoint - extend it with your own data endpoints
• See README.md for detailed instructions