ai tool
proof of concept
STATUS: Just started
This page updates you on our progress
Training City Tegucigalpa
"caminante, no hay camino: se hace camino al andar"
Antonio Machado
Artificial intelligence tool at the service of
vulnerable communities for climate resiliency
in secondary fast urbanizing urban regions of the Global South
Product is a public flooding risk heat map for Resiliency Planning
location prediction of vulnerable communities
provides economical & automatic updates
communities complement mapping on the ground (COVID open)
what WE are doing
Where
Honduras
-
Location of our Proof of Concept is Honduras: our training city is Tegucigalpa and our Testing City is TBD. Although Tegucigalpa is not fast urbanizing, it presents myriad climate issues and poverty
-
Additionally, given that Hurricane Mitch struck in 1998, Tegucigalpa was center o f several studies that will be used to train for and to validate results
-
We started developing this AI tool concept pre COVID. Now, COVID highlights the importance of using remote sensing tools
questions
More with less
-
Can we, with 20% of info, achieve 80% of certainty using Machine Learning?
-
Can we provide a userfriendly system that delivers fast and economical risk-heat- maps and its updates for secondary/tertiary cities?
-
What does it take to adapt algorithms to other regions?
TECHnology
Deep neural networks
-
Training City: Tegucigalpa, Honduras
-
Testing City: TBD
-
Data Sources: geospatial imagery (satellite, DEM) and open georeferenced datasets. We derived analytics-ready layers from these datasets such as topography, and distance to point of interest.
-
Methods: Convolutional neural networks (U-net architecture) for classification and prediction for vulnerable communities and hybrid models for flooding susceptibility.