Please use this identifier to cite or link to this item:
|Title:||Software-aided disaster risk management with Geolytics.AI||Authors:||Bhatia, Ritik||Keywords:||Engineering::Computer science and engineering::Software
Engineering::Computer science and engineering::Computing methodologies
|Issue Date:||2022||Publisher:||Nanyang Technological University||Source:||Bhatia, R. (2022). Software-aided disaster risk management with Geolytics.AI. Final Year Project (FYP), Nanyang Technological University, Singapore. https://hdl.handle.net/10356/156402||Abstract:||Accelerated industrialization has resulted in an increase in the frequency of natural disasters. Although innovations like satellite imagery have helped save several lives, there are still many shortcomings in the Disaster Risk Management (DRM) industry – 57% countries do not have DRM strategies and majority of the process is still manual. Further, lack of data verification and Community Engagement have resulted in thousands of avoidable deaths and millions in economic losses every year. The purpose of this project was to build a software solution to address the distinct lack of technology in the DRM sector. The aim was to create a technological powerhouse that leveraged the power of cutting-edge innovation such as Artificial Intelligence, automated assessments and cross-validation pipelines to provide a single, integrated platform that could address several shortcomings mentioned earlier. Specific pain points and requirements elicitation was done through detailed interviews with several organizations like National Disaster Management Authority of India and Agency of Science Research and Technology (Singapore) etc, direct interactions with ground troops and reaction forces, online surveys on social media groups as well as experts in the Asian School of Environment, NTU. This was followed by asking the interviewees on their expectations from a DRM software solution. After requirements elicitation, Geolytics.AI’s features were designed to address all the pain points that we were made aware of. For disaster detection, the lack of Community Engagement was addressed by using data from a live stream of tweets though citizen science. Latest government information aimed at plugging the communication gap between authorities and citizens, and disaster classification used Deep Learning models to validate incoming information. The design and development process followed an Iterative Development Model – once a service was developed, it was given to beta testers to gather feedback and work on improvements until a Minimum Viable Product was developed. Upon successful local development, all the features were deployed to cloud services (Azure, Heroku etc) for global, public access. The deployed version of the software was then demonstrated to several organizations and new beta testers to ensure that Geolytics.AI had a production-level quality. The results were positive and encouraging as all the organizations were interested in the product. They hailed the significant use of emerging technologies such as Machine Learning and automated services to power most of the offerings. A major portion of the feedback commended the thought that went into leveraging the power of technology to build disaster-resilient communities by, for the first time, actively involving citizens in the process. Geolytics.AI has been submitted for the Microsoft Imagine Cup 2022 and the Google Developers Solution Challenge 2022 owing to its strong business case. We are already in talks with organizations like NDMA, A* STAR and AI for Humanity to gather more feedback and setup potential collaborations to innovate further. Given its potential, there are several future extensions involving scalability and new features that will make it a more powerful application. These extensions can make Geolytics.AI an important addition to DRM strategies, capable of saving thousands of more lives in the future and enabling authorities to concentrate evacuation efforts.||URI:||https://hdl.handle.net/10356/156402||Fulltext Permission:||restricted||Fulltext Availability:||With Fulltext|
|Appears in Collections:||SCSE Student Reports (FYP/IA/PA/PI)|
Updated on Jun 27, 2022
Updated on Jun 27, 2022
Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.