May & June Floods 2015

What

• Case study and walk through of some of the features I created as part of our MOVES application in response to the flooding of May/June 2015

Remember the earlier talk?

Refresher

• Data flow
  • scraped data from sources and included it in application as geojson
    • river gauge data
    • qpf/qpe
  • Tx CAP photography of the areas impacted
  • esri base maps & some custom tiles
  • UAV data
  • custom satellite data
  • arcgis server rest endpoint (vs esri leaflet plugin)
  • other wms layers
    • google imagery service
    • army corp of engineers levees layer
    • radar data

Impressive?

• I know I'm not the most impartial person to ask but I think so...
• So lets take a look at some of this

How about that ArcGIS Server Rest API example?

Ok let's circle back around

Interesting but...

• as we added more and more data
• the app became slower... and slower

Why?

• because we were hanging more and more geojson files off of the application
• and this is still a browser based app!
• within those constraints it is impossible not to run into these sorts of slow downs
  • we had hundreds...
  • and then thousands of spatial objects that were being loaded into the app
• also, for the plane based imagery, the archive grew and grew

Solution?

Refactoring the app will fix performance!

More!

• nodejs
• pg and pg-query
• restify

Example

UT Dashboard we created

• started out with the same initial model as the MOVES app
• quickly became apparent that it didn't scale

Results

• after implementing the NodeJS and PostGres solution for the application it is both performant and scalable
• this is the model for moving forward with MOVES application