Time is a free yet rare resource. Every second matters, even more so if there is a life threatening emergency. Being prepared is the only key! In 2017 alone 335 natural disasters affected over 95.6 million people, killing an additional 9,697 and costing a total of US $335 billion. This loss could have been averted had the populace been more aware of the dangers around them and how to mitigate them. Keeping this in mind, we present ## TARAK a web app which will not only alert the user about any possible mishappenings around them, but also provide them with the required support and strategies to survive, being the true 'taarak'(sanskrit for 'Protector'). Apart from predicting posssible Natural disasters , it will also try to garner funds for ongoing major disater mitigation.
Our app will crunch in basic data from APIs across the web, geo-tagged for locations specified by the users and certain region around them. We then feed this data into pre-defined probability functions, derived by a neural net model trained on the enviornmental situations before a particular phenomenon, to obtain a probability and severity of the particular event amongst a pre-defined set of potentially dissastorous events like -but not limited to- landslides, snow-storms, Floods, Sand storms, Cyclones,Tornadoes, Tides etc. Events with low severity and probabaility index are ignored. For the rest, the probability and severity data is yet again fed to a weight function derieved by another neural network, quantifying the 'impact' of the event on the user, determined by the distance of occurrence, frequency of the event, severity, Buzz on social media, Warnings from relevant sources(Given High Priority), Casualities caused over the years, Economic loss over the past years(locally given high weight; global data also included). The events are hence indexed according to there relative weights and displayed on the user's feed. If a certain event exceeds a threshhold weight, the user is notified through a message and app notification, providing links to relevant authorities and Disaster Mitigation/ Safe-keeping techniques. A SOS button will be provided to help the users in need.
-
Estimating probable disastorous events by analyzing weather patterns.
-
A State of the Art indexing algorithm to arrange events in the order of impact for the user.
-
Area setting system, the user can set an area on the app which he frequently visits ,by setting some parameters like area location ,its size(radius).,etc. This setting will give the user daily updates on the area of his concern and ensure the safety of his loved ones.
-
Notification Based on Profile: For users in a specific profession, if the profession involves dealing with situations exemplifyinng losses from a certain event, a more sensitive notification alert can be chosen for.
Creators: Priyanshu Gupta(priyanshu@cse.iitk.ac.in); Karan Vaish(karanv@iitk.ac.in) ; Triptesh Biswas(triptesh@iitk.ac.in)