Improving Natural Disaster and Water Resource Management using Natural Language Processing

Challenge completed!

WEO is a startup with the mission to help build a more sustainable future for all by improving Water Resource Management (WRM) using data from space. 

In this two-month Omdena Challenge, a team of 50 technology changemakers collaborated to build a tool to better understand the effects of certain natural calamities (e.g. drought or flood evaluations) and improve water resource management

The problem

The goal of this project is to improve WRM and to ensure sufficient water of adequate quality for drinking water and sanitation services, food production, energy generation, and industry while safeguarding appropriate water for environmental flows and maintaining healthy water-dependent ecosystems. These outcomes become more critical across the backdrop of our rapidly growing population, increasing urbanization, changing dietary habits, and uncertainties in climatic changes.

By building a tool in which the project partner WEO can query mission-critical information from the internet, the development of innovative solutions for WRM can be greatly accelerated.

The project outcomes

Within eight weeks, Omdena’s team built a multifaceted data solution in which the user is able to define the topic of interest, the area of interest, and the time limitation for which information should be gathered. 

For an illustration here are a few examples of query topics:

• Flood evaluation: all information linked to flooding hazards occurring during and after a flood event e.g. rainfall sensors, water level information, people sharing complaints about flooding, pictures of flooded private and public spaces, and official platforms publishing flood evaluation data.
• Drought evaluation: information indicating occurring droughts e.g. plant stress/ health, agricultural losses, water level changes in surface waters, public measures are taken to ratify water usage, irrigation
• Landcover mapping: e.g. pictures are taken from and at the area of interest, information on crop planning, information on e.g. green roof installations, installations of previous pavements
• Urban climate: temperature sensors in and outside of buildings, information from heating/cooling devices, perceptions of urban heat shared on social media/blogs, etc.

Lastly, the team developed an effective way of storing the queried information/data and evaluating the data to gain certain statistical information e.g. number of images retrieved.