DisasterPrep

Inspiration

Across the world, natural disasters are occurring more frequently and with greater intensity. With the world changing constantly due to climate change caused by unsustainable practices, we are seeing more and more natural events out of balance, causing destruction and affecting our own lives.

Disaster Risk Reduction, or DRR, is a critical part of sustainable development. There are multiple parts to DRR:

Awareness - Knowledge of sustainable practices and the effects of unsustainable ones.
Education - Learning how to stay safe in natural disaster events.
Resilience - Being able to put skills and knowledge in practice to survive in the changing world

Children make up more than 25% of the U.S population. Yet, most platforms that try to address the issue of disaster risk reduction do not consider children in their planning. Most learning methods offer articles or scientific videos complicated by jargon, which even many adults may not understand. Is a 10 year old going to read a scientific article and remember anything about staying safe? Most apps offer solely one way to address disaster risk reduction.

Our inspiration for DisasterPrep was to create the complete disaster safety platform: Multimodal learning - visual, conversational, textual, and virtually kinesthetic learning.

Our platform addresses every part of Disaster Risk Reduction in multiple ways, fit for any age and any mentality. Multimodal learning is the key to addressing risk reduction and creating a more resilient, sustainable world.

What it does

DisasterPrep focuses on disaster risk reduction. Our app offers a multitude of features making it accessible to anyone, and fit for everyone.

1. Realistic 3D Physics Simulations. DisasterPrep offer 3D simulations of natural disasters, for the visual learner. Built using Unity, these simulations offer a visual experience to learn about natural disasters. There are two main options in these simulations - Experimentation Mode, and Learning Mode. In experimentation mode, the user can customize their simulation, see what would happen in different scenarios. For example, they can change the magnitude of an earthquake to see what would realistically happen. In learning mode, users are met with various interactive question-answers, so they gain the knowledge necessary for different situations for that natural disaster.

2. DisasterPal DisasterPal is your personal ML and NLU assistant, built to answer any questions you have about natural disaster safety. DisasterPal is for the conversational learner, and is capable of answer queries and providing information about various natural disaster events, safety, preparation, and so on.

3. Green Guide GreenGuide is yet another ML chatbot, to conversationally educate you about sustainability practices and how they relate to disaster risk reduction. Ask away about how to help the environment, how unsustainable practices affect our planet and lead to natural disaster events, and how you personally can create a more green, safe world for the future.

4. Information Pages Information pages are for the textual learner. While not everyone learns best through text, there are of course those who prefer reading to gain knowledge. Information pages contain thorough, easily accessible information about sustainability, disaster risk reduction, natural disaster events, and how to practice sustainable practices and stay safe and resilient.

5. Clean, Intuitive UI UI is of course a valuable part of any platform. Our UI is based on simplicity and is easy to understand. With children being a major part of our target audience, we use absolutely no complex jargon in finding information. Everything is reachable within only one or two clicks, and it is completely obvious where any information is located, whether our 3D simulations, chatbots, or information pages. Everything is even color-coded to make the experience all the more clean and intuitive.

How we built it

Our main app is created with React Native and JS. From this app, a user can access any information they need or want through the sidebar, such as the simulations, chatbots, info pages, settings, and contact pages.

The 3D simulations are built using C# in the Unity Engine. Each natural disaster event includes its own simulation with both Experimentation and Learning modes, and the user can choose either one and use the platform how they want to or need to at any point. All of the models used in the simulations were made by us using the Blender software. The simulations are hosted using HTML/CSS and WebGL.

The chatbots are built using ML and NLU in Python, hosted on Render and are accessible through the main app.
Firebase is used for contact us queries and user settings, to allow them to customize the app and tailor it to their liking.

Challenges we ran into

With so many features, DisasterPrep was often a challenge to build. One of the main challenges came from integrating the simulations and chatbots into the main app.

Integrating Unity especially into React Native was difficult as there weren't many options. After a while of looking at different options and testing them out, we were finally able to create a one-button click solution to reaching the simulations, by building for WebGL and implementing mobile touch controls.

Integrating the chatbots was also difficult in some ways. The packages used in React-Native were not directly compatible with the return format of our chatbots hosted on Render. It took a good amount of testing and trials to get it working, but it is certainly a core feature of our platform and we are glad we were able to implement it.

Accomplishments that we're proud of

Prior to building this app, we didn't have very much experience with many parts of it, particularly with the 3D aspects. We are incredibly satisfied to have been able to learn and implement 3D simulations for each of the natural disasters so our platform can provide a visual and tangible learning experience. We are also very happy with how our entire platform turned out. The UI is very clean and responsive, and the chatbots have been optimized to fit with the frontend and provide meaningful responses.

What we learned

We learned that putting together so many features can be difficult but certainly not impossible. We learned how to work with new software and will certainly use these skills in the future. Starting this project, it seemed very difficult and at many points we thought we had reached an impossible wall, but in the end there was always some solution. We will definitely continue improving upon the platform and making it even more helpful in the future. .

What's next for DisasterPrep

We have many plans for DisasterPrep in the future. Some more features we could add include integration with early warning system, creating and communicating through groups with friends and family, additional simulations, AI image recognition features to help with practicing sustainability, and so much more. There's always more to be created, with any app or platform, and we certainly plan to do so with DisasterPrep.