Support for earthquake response modeling

Contract

This is a UNV contract. More about UNV contracts.

IICPSD’s SDG AI Lab initiative (sdgailab.org) requires assistance in developing and implementing a model for earthquake safety path finding. As stated in the Task section, we are aiming to first replicate an existing approach. Please find additional information about it here: https://omdena.com/blog/machine-learning-earthquake . Our partners will be leading the decision-making process and set goals and targets to the team of volunteers to achieve the required outputs. The Lab provides in-house expertise, resources, and research support to various UNDP projects to mainstream digital transformation to development problems and to contribute to the achievement of the Sustainable Development Goals.

The online volunteers will participate in developing methods to build a model which evaluates earthquake evacuation paths and routing. The volunteers will work under SDG AI Lab supervision. The main tasks will be: 1) Develop the capacity to implement an existing model to assess earthquake safe routing and 2) Test and enhance the existing method for a) new areas of interest b) new metrics chosen to represent safety c) new data sources or algorithms to improve the accuracy. The core activities involve the detection of building footprints based on high resolution satellite images of buildings. The extraction of street network data and its representation in a graph like format to allow path finding computations. Calculating distance matrixes based on geographic information. The results of the project should be implemented and well documented in GitHub and also be summarized in a ‘Medium Article’.

Bachelor’s (or higher) in Computer Science, especially in Machine Learning (CNN’s, U-Nets, Pytorch, TensorFlow or similar), Computer Vision, Geomatics, Geography, Statistics, or related field with strong interest in Data Science; Understanding and knowledge of analytical techniques such as predictive analytics, event detection / prediction, data visualization, and experience machine learning modeling is a plus; Applied knowledge of programming languages, such as Python are required; Experience in working with geodata.