we will divide the are acording to the prone disaster
Based on the frequency and impact of a particular disaster, regions will be divided according to which disaster they are prone to . Features of a particular disaster will be used to train a machine learning model based on cumulative result of several algorithms to enhance accuracy.
Additional features like weather and climate change will be factored in.
Features (information about weather) will be periodically collected from meteorological departments and local labs.
Model will be based on cloud services like azure and will periodically input new features; regularly updating warning systems. It will update more frequently in regions where probability increases beyond a certain level. Based on the prediction, we would categorise the regions into highly prone, moderately prone and less prone to have a disaster. All notification will be provided in app or website
Precaution and safety measures will be taken according to the zone like we can notify highly prone area having higher population with red alert and inform them about safety measure they can take at that instant like in highly populated area we can tell them where to go with the help of google maps in distributed way so there will be no jams or stampede. We can also inform on social media platform to alert higher target and also various rescue teams can be informed to be ready.