TerraStory

Inspiration

The idea for this app was inspired by it's data. Recently, a team of scientists from the National Center for Atmospheric Research compiled a dataset with the latest future climate projections from ten of the best climate models. They compiled this dataset with Esri in mind; for the purpose of enhancing ArcGIS products and educational efforts related to climate science.

This and other climate datasets are large, multidimensional, and multivariate, and beckons a need to also understand GIS science. There is thus a steep learning curve with climate science, adding barriers for the majority of the population to access and analyze climate data. As a result, much of the climate debate is ripe with misinformation.

Our solution is TerraStory, an Esri Web App that provides an environment for users to have a completely new experience with climate data and other complex datasets.

What it does

By combining natural language processing and GIS, users can ask questions about climate change using verbal or text queries without scientific jargon. TerraStory then builds a query and performs a statistic analysis to compute the result. Rather than dumbing down the answer, TerraStory will then walk the user through the complexities of the result in an intuitive way using map visuals, statistics graphs, and written descriptions defining related scientific terms.

We envision our product as scalable to other Esri data exploration products including Insights and Storymaps.

How we built it

We first wrote and ran Python scripts calling the ArcPy module to process the climate datasets. We then used ArcGIS Pro to render the raster image layers, and serve the data.

The TerraStory app was built using ArcGIS Javascript API (see the GitHub link below for more information).

Challenges we ran into

Given the scope of the Hackathon, we could not use the entire global climate dataset for our analyses. We chose instead to focus the geospatial extent to California. We also chose to one of the four future projected climate scenarios and one future time period. The current app is set to answer where the driest or hottest locations will be in July in the near future (2020-2039) under the worst case scenario. Our intent is to provide a prototype application designed for an investigation of places that could experience conditions that enhance fire risks.

Accomplishments that we're proud of

The app can answer the text or verbal questions "Where will it be the hottest in July?" or "Where will it be the driest in July?". Go ahead, ask TerraStory for yourself!

We also designed TerraStory with the capacity to incorporate multiple datasets of different data types. In our current application, we include the CDC's Social Vulnerbaility Index (SVI) - 2014 map that was curated and dissemated on the Living Atlas. With this addition, user's can click in and around a region of interest and identify populations that might be more vulnerable to potential risks faced as a result of the identified climate change.

What's next for TerraStory

We have many ideas for the next chapters of TerraStory! We would like to include the full global extent of the climate data, for all four future climate timeperiods and under all four climate scenarios. We would like to add interface features that would further enable users to interact and explore the different timeperiods and scenarios.

Ultimately we would love for TerraStory to be incorporated with other complex data applications and with other Esri data exploration products.

Built With

• arcgis-js-api
• javascript
• natural-language-processing
• node.js