Bootcamp briefing special 2021

Initial situation

Floodings from rivers, lakes and at seashores are a worldwide problem for people living near them. To solve the task of monitoring water levels, sensors were installed, and water level data is published to people living in endangered areas. 

PegelAlarm/FloodAlert manages and publishes water levels from about 30,000 water monitoring stations worldwide. It notifies people about water levels that reach user-specific thresholds and thus help people living near water bodies in almost real-time.

Two problems remain

  1. Environmental factors like wind, rain, trees etc. affect the water level sensors and lead to invalid sensor data which are called “outliers”. Since there are no automated checks, these outliers are currently being published through our system. People get scared and fear for their lives if this happens at a sensor near their home.

  2. Forecasts of water levels – which are only partly contained in our system currently – would improve the quality of life of people who live in endangered areas. Forecasts would allow them to react in advance and give them more time to bring their family and belongings to save places.

Task

We are looking for developers, data enthusiasts and companies to join us in developing new methods to 

  • Detect outliers from the latest incoming sensor data to eliminate them and prevent publication. And/or…

  • Calculate reliable forecasts which are then integrated into the PegelAlarm/FloodAlert system and published to users.

Forecasts should be generated for at least 8 hours in advance to be helpful. Metrics like Mean Squared Error (MSE) will be used to verify the quality of your forecast model. Feel free to use AI, statistical methods, or other innovative techniques to solve this challenge. The quality of forecasts will improve if you incorporate not only PegelAlarm/FloodAlert’s historic water levels but also weather data e.g., rain from external data providers like https://openweathermap.org.

perspectives

We provide data access to our water level database (through API) and share technology expertise as part of the development process. We are thrilled to help you and try to answer any questions. If the quality of your service can add value to our system, we would love to integrate it and join forces for a new common business model.

Video introduction

Next steps

Get an API token from Johannes Strassmayr to access our API. You’ll find Python examples and API-documentation to query historic water levels at our GitHub page at https://github.com/SOBOS-GmbH/pegelalarm_public_pas_doc.

Get in touch and let’s improve the life of people worldwide.

Contact

The main contact person for the project on the part of SOBOS PegelAlarm is Johannes Strassmayr.

PegelAlarm by SOBOS

PegelAlarm developed by SOBOS is a water level and flood warning app, that informs you about water levels from about 30,000 water monitoring stations worldwide. It notifies you before your individual water level threshold is reached to allow you to react immediately.

More about the PegelAlarm by SOBOS

Current main briefings 2021

Additional Future Tech Bootcamp briefing specials 2021

BRIEFINGS OF PREVIOUS YEARS