Data scraped by https://github.com/simonw/irma-scrapers
The most interesting output of this repo is its commit history, which shows when the various scraping sources have been updated and what changed: https://github.com/simonw/irma-scraped-data/commits/master
Some of these changes are published to the Irma response Slack, in the #shelter_scraper_data channel. https://irma-response-slack.herokuapp.com/
Data scraped from the table on http://www.floridadisaster.org/shelters/summary.aspx
This one is a bit more complicated: we retrieve the shelters from http://www.floridadisaster.org/shelters/summary.aspx and then run a simple location comparison against the list on https://irma-api.herokuapp.com/shelters - then we generate a list of shelters from www.floridadisaster.org that appear to be missing and save that list as florida-shelters-missing.json.
We skip any shelters which have been mentioned in a comment on simonw#2 - so you can help process the list by either adding missing shelters or by posting comments about shelters that are incorrectly shown as missing.
Documentation here: https://gis.fema.gov/REST/services/NSS/OpenShelters/MapServer
Shelter data from https://google.org/crisismap/2017-irma
Outage data from https://www.sceg.com/outages-emergencies/power-outages/outage-map
Data from http://www.pascocountyfl.net/index.aspx?NID=2816
Data from http://www.hernandocounty.us/em/shelter-information
South Carolina shelters data from http://scemd.org/shelters-rss
Active shelter data from https://gema-soc.maps.arcgis.com/apps/webappviewer/index.html?id=279ef7cfc1da45edb640723c12b02b18
Animal shelter data from https://gema-soc.maps.arcgis.com/apps/webappviewer/index.html?id=279ef7cfc1da45edb640723c12b02b18
Data from http://www.ledgerdata.com/hurricane-guide/shelter/
Data pulled from https://zeemaps.com/map?group=2682928
Data from https://irma-api.herokuapp.com/api/v1/shelters
This file is derived from https://irma-api.herokuapp.com/api/v1/shelters - it shows currently detected potential duplicate shelters, based on a comparison of their latitude/longitude. Shelters within 19m of each other are considered potential duplicates.
This allows us to follow the commit history of that file at https://github.com/simonw/irma-scraped-data/commits/master/irma-shelters-dupes.json (also published to Slack) to see when someone adds a potential duplicate, or fixes one. The file also shows shelters that do not have a valid latitude/longitude and/or address.
Power outage data from https://www.fplmaps.com/ - https://www.fplmaps.com/customer/outage/CountyOutages.json
Shaped like this:
"charlotte": {
"numberofaffected": 20400,
"numberofaccounts": 114100,
"numberofrestored": 4670,
"name": "Charlotte",
"numberofoutages": 15730
}
Power outage data from https://www.fplmaps.com/ - https://www.fplmaps.com/data/storm-outages.js
Shaped like this:
{
"Customers Out": "8",
"County Name": "Baker",
"Customers Served": "5,380"
}