According to a recent UN report, India has lost around $20 billion in the last 20 years from 1998 to 2017. On the other hand, the global economic loss is worth $2.9 trillion and around 1.3 million people lost their lives due to the natural calamities. Tropical cyclones are defined as weather systems in which winds equal or exceed gale force which is around 62 kmph. With 40% of the total population of India living within 100 km of coastline, our nation is very prone to cyclone activity.
What do we aim? During any calamity, humans are concerned mostly about 2 things: life and assets. Most of the valuable assets are fixed entities and can not be moved easily during disasters. Critical Infrastructure Management involves the study and analysis of models, methods of the real-world infrastructure (considered as important for a nation’s economy) to reduce or avoid the consequences of the destruction. We envisage to build a real-time cyclone tracking and warning system for citizens based on the critical infrastructure management. The central idea is to alert the citizens in advance and keep them posted of the status and location of the cyclone as it approaches in real-time, thereby attempting to save lives and property.
The plan...
We aim to use a Complex Event Processing and Deep Learning based approach to maximize the accuracy of our analysis by running computationally intensive algorithms over real-time sensor data-streams. The Complex Event Processing engine would continuously monitor the sensor data streams to filter and identify anomalies indicating potentially hazardous weather events. The identified anomalies would then be fed to a deep learning model to further refine our analysis. The results would then be conveyed to the citizens in the form of alerts and warnings indicating the status and the location of the cyclone in real-time.
How can we use the project?
The National Cyclone Risk Mitigation Project (https://ncrmp.gov.in/dos-donts-cyclone) by the Government of India highlights some points for the residents to minimize the catastrophic loss and casualties. Our project will help the users to make efficient and sensible decisions based on Geographic Information Systems (GIS) analysis. The users can setup our project and get the real-time scientific visualizations about impact, velocity and location of the upcoming disaster.
Datasets:
We plan to use following datasets for study but we might add some more resources if required:
- National Institute of Informatics Japan, Digital Typhoon: http://agora.ex.nii.ac.jp/digital-typhoon/links.html.en
- Stormwatch by Weather.org: https://www.weather.org/stormwatch/
- ISRO Disaster Management Support Programme, NRSC Decision Support Centre: http://bhuvan-noeda.nrsc.gov.in/disaster/disaster/disaster.php?id=cyclone
On what we would build?
- For Complex Event Processing: Microsoft Complex Event Processing Engine, Azure Stream Analytics
- For generating streaming data: Apache Kafka
- For building Machine learning model and scientific Visualization: Microsoft Cognitive toolkit and Azure Cloud services
- GIS Analysis: Open-Source available applications as Qgis