Space Apps Challenge 2018

FUTURE SUTURE
“A stitch in time”

A mixed reality experience center that combines machine learning and Virtual Reality to provide a gamified experience to facilitate climate education.

Highlight: NASA Space Apps Challenge Seattle Hackathon winner for Most Inspirational Hack.

Future Suture is the brainchild of Team Stardust Explorers (comprised of Proshonjit Mitra, Prithvi Shetty, Varun Panicker, Bhagyashri Nivdunge, and Pratik Mulchandani). This team has a good mix of Data Scientists, Developers, and a UX Designer, and keeping that context in mind, our team chose to accept the GLOBE-ally challenge of the NASA Space Apps Challenge competition. In this challenge, our goal was to effectively communicate insights using NASA data.

Brief Description:

Future Suture is a virtual experience center that combines Virtual Reality and machine learning to simulate climate and environmental conditions. By using NASA data about climate conditions, create machine learning models to visualize how the future would look like based upon a given set of variables. Variety of actions would result in a variety of future possibilities. However, each city would have a different picture that can be painted. For example, users in Seattle can see the ice caps of Mt Rainier melt causing flash floods, or see the waterfront submerging a portion of downtown Seattle. These futures can be simulated in a VR+4D environment and can be experience in the form of a game that basically puts the onus of the future on the user’s actions.

For this project, we created some predictive models to back our future simulations and prototyped the gameplay in VR form. Our team won the ‘Most Inspirational Hack’ award and also became one of two teams to make it to the Global round of the competition and got a chance to represent the city of Seattle in a global competition.

1. Motivation

To increase awareness about environmental issues and encourage people like Bob to act upon it.

• Climate Change is already having significant and harmful effects on our communities, our health, and our climate. Sea level rise is accelerating. The number of large wildfires is growing. Dangerous heat waves are becoming more common. Extreme storm events are increasing in many areas. More severe droughts are occurring in others.
• We must take immediate action to address global warming or these consequences will continue to intensify, grow ever more costly, and increasingly affect the entire planet—including you, your community, and your family.
• Global warming and climate change awareness is the first step towards the sustainable future. Our motivation is to make everyone think globally and act locally towards the sustainable future.
• If the future generation is made aware of the circumstances about the current situation, ‘Future Suture’ can trigger a chain reaction and alert everyone about the hazards of harmful human actions on environment.

2. Aim:

• Local centres:

The goal is to establish multiple ‘future suture’ centres across the globe.

1.Providing a future simulation of the region using machine learning:

Imagine you live in Seattle, you will be shown the future condition of Seattle based on the actions you perform in the game.

2. We will provide futuristic measures to protect the environment catered to the specific region.

For example, a person visiting the ‘Future Suture’ centre in Dubai might be encouraged to use solar energy as an alternative source of energy. A person visiting the ‘Future Suture’ centre in Seattle might be told to use ‘hydro-electric’ power as an alternative source of energy.

This is how the game looks like :

We will highly recommend to watch the prototype of ‘Future Suture’ to experience and get to know the exact motive.

Youtube link : https://www.youtube.com/watch?v=chDnJkFTx2s

• Climatic records and monitoring:

Climatic variables such as temperature, pressure (surface), precipitation, water vapour (surface), surface wind speed, surface radiation budget are considered while collecting data.

• Temperature forecast:

The ML model forecasts temperature levels for us to model our VR system on. The pictures in the surrounding areas around the person will reflect the effects of these increasing temperature. These increases will not only be in terms of the heat or water levels but will also include the effects on the local flora and fauna.

3. How will it work?

1. The user’s location is used as the default region. We will use this to obtain specific climate data to that region.
2. We use the region specific data and feed it to our machine learning model. This data can be both, historical as well as live data from cloud data sources.
3. The model predicts some environmental conditions such as temperature and sea levels in the surrounding area.
4. The climate predictions obtained are fed into our virtual reality system to display visual representations of the person’s surroundings.
5. Here, the person is taken through a series of screens where she can make choices that will reflect directly in her VR system.
6. If the person fails, we provide them with recommendations to reduce their carbon footprint and avoid the undesirable future.

4. Details

Prediction and Forecasting Process

• Temperatures for the next 50 years have been predicted using historical data forecasting. We have considered variables such as Humidity, Air Density, Water Vapour concentration, and Water vapour deficit. We were able to predict the temperatures using these attributes with an accuracy of 97% on our test data.
• We used regression with least squares for this purpose considering the number of data attributes at this point was not large and effective predictions could be made without the use of more complex algorithms. However, if we are to include additional data such as readings from weather satellites, glacier temperatures and natural phenomena such as storms etc. we will have to turn to more complex decision driving algorithms.
• Our current model is based on batch learning due to the strictly historical nature of the data source but can be easily scaled up to an online learning model to train incrementally to constantly updated data sources such as satellite data.

Machine learning model

5. Impact:

• For local residents:

Many people like Bob refute the effects of climate change because they cannot see them. They feel that climate chain is abstract and if at all true will not affect them directly. We want them to understand what effects adverse changes in global temperatures will have on their lives by showing them their surrounds post these changes. The residents of an area will be able to experience the future based on the predictions made on the available climate data and learn how to take the required steps to make informed decisions and the impact of their actions. The objective of our initiative is not just to spread awareness but also to teach people how to reduce their carbon footprint.

• For the government:

The government will find a far more popular way than usual to reach as many people as possible and make a lasting impact at the younger generations by driving education through experiential learning. The app or game could be built inside a VR headset that can be distributed to schools and libraries for the general populace to use. The interactive game can make for a great learning solution for global warming in schools.

GitHub: https://github.com/shettyprithvi/NASA_Hackathon