The effects of climate change are an inevitable truth. Climate change has already and will continue to noticeably affect many critical aspects of day to day life of many people around the world. At this point, it is necessary to consider the future impacts of climate change in order to have certainty in the security and safety of your geographical location. Our team decided to build a web application that interprets Google Cloud Platform’s Earth Engine datasets captured by satellites to teach people about the impacts climate change may have on their everyday life, if things go unchanged. Such a tool can also improve the health of those considering environmental factors when making a decision of where to live in the future. Additionally, this app can help individuals make plans for future decades, if it is necessary for them to relocate for their safety. Moreover, applications like these highlight the tremendous impact and value space technologies like satellites have on everyday life.

The web application takes two entries, a location and a year date. Using these two points, it obtains historical data regarding this region from multiple datasets provided by Google’s Earth Engine (Google Cloud Platform), including precipitation, temperature, changes in local water level, natural disaster threat levels, and other factors like atmospheric analysis of certain compounds. After obtaining data for as far back as the information goes, the application attempts to make a future analysis for up until the user’s specified year date. The application then returns various data charts and an index of how livable the specified area will be in said year.

We ran the React application on Flask to get user queries from another IP address and respond with information. We used React.js to access the information from the website as well as display the requested information. We then searched Earth Engine’s repositories for datasets we could use to build our project. We used various Python libraries to analyze the data extracted from Earth Engine, including statsmodels to train the machine learning models, and matlib to generate and send images to the frontend.

The foremost issue was determining how to interpret the data from Google’s Earth Engine. There is a vast repository of data regarding hundreds of different aspects of Earth, and this data is stored as a wide variety of differing variables and types. The documentation for each dataset is not as comprehensive as we hoped. We had to learn how to interpret scientific data and then use technology to analyze and accurately extrapolate said data. We also had to learn how to efficiently parse through images to cut down the large amount of time it takes to train the model. In addition, the team had to learn new tools such as Flask and server hosting, both of which are essential to the project. While we would have loved to demonstrate analysis of various tools like pixel analysis of vegetation and atmospheric analysis of various compounds, we realized that the processing time was too great for this demonstration and that a creation of a truly accurate model would take some more time than we have for this event.

We are proud of our ability to have created a web tool that not only will help people make a decision about where to live, but also educate people about the real effects of climate change and about the usefulness of space technologies like satellite sensors and imagery. We are also proud to have accomplished so much in the time period allotted, especially for a team that consisted of members who did not have experience in the technologies used.

Our team learned how to efficiently analyze large datasets for useful information and trends, in addition to training a model to make accurate analysis of said data. We learned how to deploy a functional web app over a server. In addition, some members of the team were introduced to libraries like React.js for the first time.

We hope to continue to add more significant features and analysis to our tool. We realized that there are truly vast amounts of information about the climate and Earth about a wide variety of useful categories. Potential features include better analysis of vegetation, droughts, and natural disasters. We would also look into stronger prediction models more accurate to the specific dataset.