High-Level Project Summary
Such an event like the Carrington Event today might cause unprecedented damage as the world has become far more dependent on electricity than it was when the Carrington Event occurred. Our solution is to make an early prediction algorithm to predict the next potential Carrington-like event and deploy it on a website as it could give humanity a warning about potentially dangerous space weather. Using solar wind data, provided by the DSCOVR spacecraft to track and follow the changes in the peak solar wind speed, we can train the model on these provided data and calculate the probability of massive solar storms. Once the probability crosses a certain threshold, it will send a warning.
Link to Final Project
Link to Project "Demo"
Detailed Project Description
Our deep learning model detect and track solar wind peaks and provide provides early warnings of the next potential Carrington-like event. It’s a neural network model, specifically LSTM trained on NASA data provided by The Deep Space Climate Observatory (DSCOVR) is the National Oceanic and Atmospheric Administration (NOAA) in the resources. It works well on the given data and gives good accuracy. Also we deployed it on a website so it will be available to:
- Subscribed companies including satellites and electricity companies that subscribe to get mails in each possible danger to take their precautions.
- People who live near the poles so they’re in much more danger than others to face the Carrington event; can join our website in which we will provide safe shelters for them in case of that event happened AND this will be funded by our profit from subscribed companies.
- Free subscription to authorities.
- Guides to people to be more aware about how to survive through the event.
- Influencers to make awareness content about Carrington event to make it viral and how to survive during it. We will provide them a section in our website to share their content.
We hope through our model and website to raise the awareness and save the most possible to save from the Earth. We hope to reach more companies and influencers to help us providing a shelter for people in danger during Carrington event.
Used tools:
· Coding languages: -Python -HTML -CSS - JavaScript
· Software: 1- Jupyter notebook 2- VScode 3- Brackets
· Libraries and Framworks: 1- Flask 2-Numpy 3- Scipy 4-Pandas 5-Matplotlib 6- sklearn 7-Tensorflow
Space Agency Data
We have used the space agency data provided by The Deep Space Climate Observatory (DSCOVR) It was built by NASA and is operated by the National Oceanic and Atmospheric Administration (NOAA). It is a Space Weather Prediction Center's principal asset for monitoring space weather and providing early warnings of solar events that could affect Earth. We used this dataset to train our model and this helped the model to have high accuracy.
The Deep Space Climate Observatory (DSCOVR) mission is now the primary source for real-time solar wind and interplanetary magnetic field data but there is one more satellite at the Sun-Earth L1 point that measures the incoming solar wind and that is the Advanced Composition Explorer. This satellite used to be the primary real-time space weather data source up until July 2016 when DSCOVR become fully operational.
Hackathon Journey
NASA Space Apps is a challengeable, but with much joy in same time. We had a lot of hard moments but we made through it and this mad us able to work under pressure. We learned to do brainstorming well. We tried a lot of productive strategies that most of them failed but finally we are here. We made good friends, connections and expanded our technical and personal network.
We chose this challenge specifically because we want to save the Earth from a danger after we attended the boot camp session about Carrington event. We discovered that Aurora is dangerous not a beautiful phenomenon, so we wanted to make people know about that and not be dazzled with just good appearance.
We are inspired of DeveOps, DataOps and Agile that make project management process through the Hackathon. We combined the forces together and worked in a way that makes us all connected at the same time. And another management approach which is “Problem, Solution, benefits” we merged them and understanded each one and that appears in our deployment to the website.
We resolved setbacks and challenges through trying more than one time to solve ourselves, then asked a mentor. We tried to calm down through intense situations and not to freak out.
We would like to thank NASA space apps for providing this nice experience which is highly organized and useful to all of us. We want to thank volunteers and mentors for helping us. Finally, we want to thank for making through hard times in these two days and managed to get the basic value from the hackathon.
References
1-https://www.researchgate.net/publication/355367493_Predicting_Solar_Flares_with_Remote_Sensing_and_Machine_Learni ng
2- https://iopscience.iop.org/article/10.3847/1538- 4357/aac81e
3- https://2022.spaceappschallenge.org/challenges/2022- challenges/carrington-event/details
4- https://scijinks.gov/what-was-the-carrington-event/
5- https://www.livescience.com/carrington-event
6- https://www.huffingtonpost.co.uk/robert-hill/using[1]machine-learning-to-2_b_18228808.html
7- https://www.aalto.fi/en/news/new-forecasting-tool-can[1]give-an-early-warning-of-solar-storms-boosts-preparedness[1]and-helps
8- https://www.swsc[1]journal.org/articles/swsc/full_html/2021/01/swsc200032/swsc 200032.html
9- https://cdaweb.gsfc.nasa.gov/pub/data/wind/mfi/mfi_h2/2022/
10- https://cdaweb.gsfc.nasa.gov/pub/data/dscovr/h0/mag/2022/
11- https://spdf.gsfc.nasa.gov/pub/software/cdf/dist/cdf38_0/
Tags
#software #AI #ML #Earth_Savers