Hi! Greetings from the team Identity Crisis. As a common reason behind projects with most joyful development phase, our project is also need inspired. Lungs of the Earth, Amazon rain forest was on fire just a few days ago and we couldn't really do much about it. Nature controls our lives in weird and exciting ways but its up to us to get up each time we fall down.

Our project is more than a visualization app for forest and other kinds of fires around the globe. Most of the wildfires are caused due to human mistakes and only a small percentage of it occurs because of our mother nature. We believe that people in our society are lacking severely in awareness about these fires which are often repercussions of their deeds.

- Our app focuses on enabling the people to check the Danger Levels of a wildfire in a specific month, in their town or area.
- It enables the users to report fire around them when it occurs.
- It has a feature to suggest the authorities about exact ideal locations where fire fighting stations and similar centers should be established so that they can reach an emergency spot as soon as possible.
- It helps the users to visualize the locations on map where there is an active fire.

Our app is using NASA's FIRMS data from MODIS to enable the users visualize active fire locations and the archival data set helped us building the machine learning model (using SVR) used in predicting the danger levels of a fire in a particular are, in a month. The danger levels predictions are dependent on the fire radiative power and the brightness data provided by the FIRMS archival data set. Clustering algorithms (tested on kmeans, random forest, DBSCAN) are used to figure out zones and sub-zones in a nation and then give the centroids of these fire prone locations as ideal points for building fire fighting infrastructure. Near real-time data of fires around the globe is used to verify the reports which users post on our platform of active fires near them (approx 3Km radius). As soon as a fire report is found to be positive the fire authorities are alerted about it.

We are totally down for working on this project in near future and we are planning to build a full fledged social media around this app where people can post their verified fire reports and can search for other such reports in their specific sub-zones.

We have used Java 12, XML, Python 3.7, Android Studio and Anaconda 3 for building this project from scratch.

GitHub URL: https://github.com/MrRobo24/nasa_space_apps_challe...
Tags: #wildfires, #spotthatfirev2, #modis , #machinelearning, #python, #datascience, #pandas, #numpy, #svr, #kmeans, #spotifire