Several disasters have been hitting the Mother Earth over the period of time. Just recently, an earthquake hit Italy. Death toll is nearly 300. Few months ago, a magnitude-7.8 earthquake hit Ecuador claiming 272 lives while affecting 2500. In October 2015, an earth quake struck India, Pakistan and Afghanistan recording nearly 8.0 on Richter Scale. In the same month, severe flooding occurred in Chennai, India, which claimed 70 lives and effected 10,000. Tōhoku earthquake and tsunami hit Japan in 2011 recording 9.0 on Richter Scale, affecting 228,863 people with insured losses US$14.5 to $34.6 billion. A decade ago, a Tsunami hit Thailand which is considered as the deadliest tsunami till date claimed over 280,000 lives.
All these struck in a split second. But the aftermath was very destructive. Even with all the latest equipment and technologies, disaster were not predicted or not predicted accurately. Japan is most susceptible to different kind of disasters. According to several sources, Japan has been trying to predict ‘earthquakes’ for a long time. It has deployed several sensors in seas, but it has not reached satisfactory accuracy level. Similarly, other countries have also failed to predict disasters.
Does that mean we should remain at the mercy of nature forever? No!!! We can use nature to ‘predict’ nature!
Several sea merchants claimed that they observed abnormal behavior of fishes before the dreadful tsunami hit Thailand. My mother observed the abnormal behavior of pet birds few minutes before earthquake struck my country in 2015. This leads to a conclusion that animals and/or birds are very good at predicting disasters. My idea is to monitor this abnormality and alert the neighborhood.
Technical Side: However, please note that the eye witnesses were just laymen. But with latest technologies and improved infrastructure, we can predict it even better. All major cities across the globe have zoos. The idea is to install sensors on animals and to monitor their behavior (just like current IoT platforms: http://www.datasciencecentral.com/profiles/blogs/can-a-cow-be-an-iot-platform and award-winning http://www.cowlar.com). Sensor would be a collar band comprising frequency detector, accelerator and GPS chips. Or otherwise, sensors may be installed in animal cages. These sensors will transmit live streaming data to the nearest regional head quarter, which houses a prediction model and a Hadoop ecosystem. Efficient data mining, with expertise in handling imbalanced data, will accurately predict the disaster. Whenever there is any uncertainty in their behavior; that is the alarm. A disaster is striking soon! Alert the neighboring area or the entire city over smart phones (just like Google Weather warning) or text messages etc. to avoid all possible life loss. This is a great application of Internet of Things (IoT), taking it to its new heights.
Profit Estimates: First year cost sums to be around $100,000 including cost of sensors, licenses, labor cost of building prediction model, marketing cost, cost of hardware, cost of office space, cost of building and maintaining alert system etc. Whereas, revenues sum to be atleast $200,000 (assuming alert system subscriptions charged at $1.00 per year to 1% of population of an average metropolitan city, starting after 8 months of initial development). First year profit shall be atleast $100,000. This shall significantly increase in subsequent years as one-time costs diminish, whereas alert system subscriptions continue to increase.
According to http://www.theguardian.com/world/2016/mar/03/indonesia-tsunami-warning-buoys-failed-when-quake-hit-officials, Indonesian buoys recently failed to predict a disaster. Several 'forecasting' computers and models have been built in the laboratories of leading universities around the globe (http://www.newyorker.com/news/news-desk/the-new-science-of-disaster-prediction). But they are either limited to just earthquake prediction, or they are still inside the laboratory for 3 years, or they are not so accurate, or the time lag is insufficient. But the best part of my idea: it can predict any disaster. Yes, it is not just limited to earthquakes, but it will predict tsunamis, heavy floods, volcanic eruptions, tornadoes, and others.
In the future, based on the on-going disaster prediction time, buildings should be designed such that all people are able to safely evacuate the building within the on-going prediction time (i.e. time interval between a disaster is predicted and when it actually hits). Moreover, behavior in animals can be investigated further as to which animal reacts to which disaster best and how much is the response time to predict each disaster.
This is a highly observed idea. It can be tested, with Classification Matrix accuracy confirmed to be much higher than accuracies of existing models. To avoid precious life loss, this is the solution!
If you want to make millions, or possibly billions, inbox me asap !