Last week was the 1-year anniversary of the deadly Montecito mudslide. We wanted to create an app to improve community safety by alerting officials of structures susceptible to damage.

We predicted and modeled potential mudslide damage using rock strength and slope steepness data from the California Department of Conservation, real-time precipitation forecasts from the National Weather Service, and 125 million building footprints generated from Microsoft's deep learning algorithm. 

What it does

SLIDE helps decision-makers prioritize closures and evacuation areas BEFORE a possible mudslide event. The app presents a map of all structures susceptible to mudslide damage and provides all addresses impacted. Officials can download the addresses and use it in Workforce to assign specific buildings to street crews. They can learn vital statistics about the population living in mudslide susceptible areas to better prioritize emergency response.

The user inputs their area of interest and runs the app. The app finds all areas that have a high landslide susceptibility index AND are in an area that is predicted to have over 1 inch of rain in the next day. It selects all the building footprints within the resulting area, creates centroids of the footprints, and reverse geocodes the centroid to find the address of each point. The app finds all Block Groups that intersect the centroids and creates an infographic showing vital emergency response information such as population, number of households without vehicles, and non-English speaking population.

How we built it

ArcGIS Pro and ArcGIS Online: Calculated percent area of hexbins with high risk for landslides using Raster to Polygon, Intersect, and Summarize Within tools. Enrich Block Groups for information most helpful to emergency response crews and officials.

ArcGIS JavaScript API: Created map component for the front-end app

Node.JS and ArcGIS Rest API: Built a web service used as a middleware service between ArcGIS Online and the front-end app. Web service finds high risk hexbins that intersect precipitation zones of >1 inch, creates centroids of building footprints within selected hexbins, and reverse geocodes centroids

Webpack: Manage components to build the front-end app

Challenges we ran into

•The lights turning off every half hour in M2 Studio. Solution: Get up and dance til' the lights turn on!

•Analyzing Microsoft's 125 million building footprints

•Converting map service to vector then addressing issue of too many vertices in the vector output

•The reverse geocoder does not always identify every address when batch geocoding.

Accomplishments that we're proud of

•Our logo

•Creating a Splash renderer from scratch and connecting it to the main webpage

•Popup template generated from another query and not a feature layer

•Converting building footprint to centroid, reverse geocoding, then outputting all addresses for download

•Data preprocessing and GIS workflow

•3D visualization of landslide susceptibility

•Incorporation of WebApp Builder widget within app to generate demographic statistics for areas of interest

What we learned

•The workflow of the view and controller framework


•We interviewed three public safety experts, Stacey Triche, Chris Ferner, and Jon Pedder, to learn what people in the field actually want in a GIS solution.

•WebApp Builder Widgets

What's next for SLIDE

•Implement direct connection with Survey 123, Dashboard, and Workforce

•Public-facing app that highlights road closures and helps people find alternate routes if their route is blocked (using the Directions Widget that sets a feature layer [mudslide] as a barrier)

•Allow public to sign up for push/text notifications of public safety warnings, road closures, etc.

•Predict depth of mud and show in 3D next to houses (e.g., 3ft = Destroyed, <1 ft = Damaged, etc.)

•Ingest Weather Underground's citizen engagement program which has access to people's personal weather sensors in their backyard

Built With

Share this project: