Today, India is one of the worst flood-affected countries in the world. India with 20% fatality rate is one of the worst flood-affected countries in the world. The Bihar state of India witnessed one of its worst floods in the monsoon months, i.e. July and August. Due to this flood, around 3000+ people lost their lives and the houses of 20 lakhs people got damaged. Due to lack of information many people lost their life and this motivated us to work on our project SAATHI- the Saviour

What it does

As it is quite clear that flood and drought are indeed important and devastating calamities which affects a large population of a demographic every year. So in this situation, if can make a system, or for large user, an app, which analyses the rainfall of the region a person is living in, and can raise alarm for the person if there might can arise a situation of flood, he/ she can take useful measures to ensure least damage to him and his property. Same thing could be used by the Indian Meteorological Department of India to plan and process these events and make sure less people are affected by the calamity.Our Project gives real-time prediction about future floods and as well as our hardware provide a easy way to measure water level.

How we built it

We built our hardware using JSNSR04T waterproof ultrasonic sensors for measuring the water level at every particular time interval, NRF24L01 for node to node and node to gate connectivity, GSM module for sending data to our server and Arduino nano will controlling all the modules and sensors.So how are we going to do it? We are going to use machine learning, deep learning and statistical learning theory to train our model to predict the calamity. We have got our data from Indian Govt Website link, it contains the monthly, annual and seasonal rainfall data of each state and their major district of India. The data is 50 years, from 1951-2000. Also we are getting instantaneous data of rainfall (visually data after each day) from Indian Meteorological Department link.The data looks like this.

Phase 1:

Simple Prediction using the above stated model. Independent analysis of the data( independent in the sense that rainfall in Uttrakhand will not have any effect on the people living in Delhi)

Phase 2:

Correlational Analysis: It is quite possible that a heavy rainfall in North can bring flood in all the areas which are in the vicinity of that area, and it is also important to analyse that part. So after making the basic model, we'll try to incorporate this aspect also in our model. 1.Training the model on previous data.

2.Taking USER's Location from the APP.

3.Running the algo at back-end.

4.Predicting the chances of flood.

5.Notifying the User.

Challenges we ran into

The biggest challenge was transmitting the hardware data and using it as real time from various stages and also we were working on TigerGraph for the first time so it was two way journey of learning and exploring for us.We saw how complex data can be visualized very easily and cleary.

Accomplishments that we're proud of

We're proud of that we have successfully able to make the entire project and provide a better solution for flood prediction with a minimum cost of 20$ which can be less if produced on large scale. We believe that this system is going to help a lot of people as well as the govt in predicting the disaster and also in reducing its impact on general people. Every year people won't have to suffer the same, and hence we'll be able to make society a little better with this.

What we learned

The most important skill we learned is to connect the microcontroller to AWS Gateway Api also how to trackel and enjoy the tigergraph overall working and the power of team work

What's next for SAATHI THE SAVIOUR

By using more accurate and a large amount of dataset we can predict more accurate results and forecasting of Flood.

Share this project: