High-Level Project Summary
POSEIDON project is a machine learning-based project that predicts when and where floods will happen. Using a machine learning model trained on NASA's satellite data of wind, soil moisture, and clouds — all of which are factors that lead to a flood — then we deployed this model to a website with a user-friendly experience that allows anyone to use whether from the government and decision-makers or researchers or normal people to know the possibility of rainfall and floods in vulnerable communities which may help them address the situation early to rescue humans there and by that our role ends with define these communities and highlighting their suffering.
Link to Project "Demo"
Link to Final Project
Detailed Project Description
How We Addressed This Challenge !
Our idea is to build a machine learning model that can predict when will be the next flood and create a website that is available and easy to deal with for the public to understand and be aware of what they should do. It depends on predicting the possibility of the flood's occurrence and gives them an alert before it . The program will calculate the probability of a flood by providing certain data as a result of monitoring satellites for NASA. The solution to that problem in a place like Sudan By directing that water to the plain lands before reaching the population places and causing losses and making use of the water to provide other areas for agriculture.
How We Developed This Project
floods have killed hundreds of thousands of people or maybe millions. They do not stop here as the flood water can carry infectious diseases and chemical hazards causing more damage and deaths. Despite knowing that floods will happen and actually some places are known with continuous floods throughout the year, yet still we can't predict them. You wait until the flood hits your home and then maybe you will get some time to scream. A lot of wasted souls and water then we should take some action. But don't worry POSEIDON is here for you.
It was necessary to use artificial intelligence to reduce these damages and protect these societal groups. Therefore, we used the python language to create a model that can predict the occurrence of a flood in a specific range, which saves the residents of that area from leaving or working to direct that water to remote places, through a website that is easy to deal with and understand. The possibility of flooding is displayed on it
What was your approach to developing this project?
The first step was to get the data required from satellite imagery from NASA’s worldview servers. The specific satellites we pulled from are NASA’s Terra (MODIS instrument) and SMAP satellites. Now the data for normal conditions and disasters are recorded (soil moisture - clouds and rain clouds condition - wind speed and direction).
In the next step we give code the data directly from satellites through servers for the navigational condition of the air and climate changes that It occurs directly, so it studies and compares it with the data previously recorded.
In the third step, we display the data resulting from the comparison on a simple and understandable website, which in turn shows the possibility of a flood in the specified range
Space Agency Data
We used earth data, and then soil moisture readings from SMAP satellites. As for the atmosphere and climatic conditions that may lead to flooding, we used JAXA to know the wind speed and the state of the clouds. As for the injustice Social and damage caused, we used (sedac) to determine the most vulnerable countries to disasters and marginalized communities.
Hackathon Journey
What problems and achievements did your team have?
Problems:
Space Apps experience was a stressful and full-of-learning journey. None of us know each other and we formed a team together and tried to think and share ideas, sometimes arguing but in the end we managed to get an idea and submit it before the deadline. After our idea has been accepted, sleepless nights start.
It was the first time for all of us to use any of NASA’ data and maybe without these hackathons none of us would have used it. We were drowning in many sources and found it hard to get relevant data to continue. Each one was responsible for something and we all developed our skills.
Achievements:
We wanted to make something with impact. We live in a quite safe community and we never experienced any floods or a thing to call disaster but watching videos and reading news about what is happening to other people was really hard. Watching someone trying to catch a robe or a piece of wood to avoid dying. Someone has to do something.
The most difficult part was finding the data and also choosing only one community to address and trying to give an example of vulnerable communities. We changed our project many times until we settle on something applicable
References
Resources :
- https://smap.jpl.nasa.gov/ (smap)
- https://earthdata.nasa.gov/learn/pathfinders/disasters/floods-data-pathfinder#tools (floods nasa)
- https://www.eorc.jaxa.jp/water/map/index.html?area=global (soil moisture sat)
- https://gportal.jaxa.jp/gpr/search?tab=0 (climate & water case)
- https://sedac.ciesin.columbia.edu/data/set/pend-gdis-1960-2018/data-download ( Socioeconomic Data )
- https://www.nature.com/articles/s41597-021-00846-6 (global dataset for floods)
Project code :
https://github.com/MoNasr0/poseidon
Videos :
Tools :
- Atom
- Adobe Photoshop
- Online tutorials
Tags
#POSEIDON #Floods #Environmental_hazard #machine-learning #AI #big-data #satellite-prediction #Enviromental_injustice
Global Judging
This project has been submitted for consideration during the Judging process.