Space Apps Challenge 2018

Explanation

NASA collects data from citizen scientists around the world who are looking at clouds, mosquito habitats, and land cover from the GLOBE observer app. These data along with NASA satellite data can be used to identify and communicate information to improve life on Earth. But existing tools and visualizations are not friendly to use and understand. We have each identified problems that can have a global impact and proposed to build an app that solves for them. We have used the data available and identified mosquito habitat cover, deforestation, solar energy adoption and the cash crop chain as few of the problems that need an innovative perspective.

Validating the idea

We brainstormed together to identify problems to solve using the available data by posing short questionnaires to fellow hackathon participants at the venue. In return, we provided feedback to their ideas and created a win-win situation for all!

We learnt a lot from our own research as well. We learnt that the data visualizations available did not clearly provide a protocol of action to take in case a situation arises.

• The hydrosphere data from GLOBE for example did not provide the probability of contacting the Zika virus in a given location.
• Solar power adoption in a tropical country like India was very minute and some basic data analysis of population cover and per-capita income showed that this was because of the large up-front cost of installation of solar panels. This also prevented most potential buyers from seeing its long-term benefits!
• We also analysed causes of deforestation to see if they could be reduced in any way. To our utter disbelief, we found out that forest fires are the most preventable form of deforestation and that there was no way to predict them accurately enough. Predicting them can easily mitigate large-scale damage!
• We read a few articles on cash crops especially export surplus ones being burnt in excess. This alarming fact throws light on supply chain optimization in order to route the crop yield better. This can feed the hungry and also fetch higher profits to the farmers.

Our solution

The problems we chose to take up spanned four key categories comprising health, nature, renewable energy and food crops. Each having a global impact if solved sufficiently well. With the time on our hands, we chose the following:

• Problem 1: Mosquito borne diseases cause millions of deaths annually
• Problem 2: Deforestation is accelerating, despite mounting efforts against it
• Problem 3: Slow rate of adoption of solar energy due to economic factors
• Problem 4: Inefficient farming practices and supply chain leading to low yield and resource wastage

Our Solutions are based on using solid data backed insights and intricate visualizations to help bring awareness and mitigate the potential damage they may cause. We have also used creative solutions based on Machine learning and deep learning models

• Mosquito Attack: Using a modified linear regression model to track Zika virus spread across the globe to predict the probability of contacting it using a GeoChart visualization
• Preventing Forest fires: Using Convolutional Neural Networks with a selu (not relu!) activation function to predict forest fires and take steps to mitigate its effects
• Adoption of Solar solutions: Using Monte Carlo simulations and curve fitting (Normalization) to visualize solar adoption using a solar power usage calculator and also provide a temporal irradiance visualization
  
• Better farming solutions: Using GeoCharts to visualize prime cash crops and mapping a profitable supply chain maximizing profits for export surplus

Project Website

The link above leads to the web app

Our Final Results

(The site did not allow multiple picture embeds. Hence the visualizations have been linked to the screenshots on our Github repo)

Project Repository

• Zika probability visualization
• Forest fires geochart
• Visualization dashboard
• Solar Calculator
• Cash crops excess distribution
• Globe 3D climate visualization

Project Slide Deck

Team

• Abhijeet Singh
• Susmit Agrawal
• Vipul Gaurav
• Vishal Vinod

Resources

Our solution would not have been possible without these resources:

• Global Imagery Browse Services (GIBS): https://earthdata.nasa.gov/gibs, provides access to over 400 satellite products.
• NASA Worldview: https://worldview.earthdata.nasa.gov/, an easy way to visualize GIBS data in your browser.
• NASA’s Earth Science Data Systems Program: https://science.nasa.gov/earth-science/earth-science-data, describes several sources of Earth science data.
• NASA Earthdata Search: https://search.earthdata.nasa.gov/, allows you to search, discover, visualize, refine, and access NASA Earth Observation data.
• GLOBE Advanced Data Access Tool: https://datasearch.globe.gov. This tool allows you to find and retrieve GLOBE data using several different search parameters.