Thread by @ChristieStClair

THREAD/ Yesterday, two years ago.

Me and five online friends watched as Twitter users pleaded for flood rescue during #Harvey. The 911 system went down. Fed, state and city officials told people to use 911, and not to use Twitter.

We saw as anxiety turned to fear, then terror.

We saw volunteers wasting critical time boating to homes based on outdated posts.

We decided to focus on supporting fed agencies by giving them actionable, accurate, timely data for search and rescue.

First up was the Google spreadsheet.

Then we coined the term #HarveySOS, and asked friends to go out and tag distress calls using that hashtag.

Another team grabbed those tagged tweets and entered them using consistent formatting.

That data fed into a map we built, so rescuers could see exactly which streets, buildings and communities needed help but couldn’t (or wouldn’t) call 911.

Interestingly, some research I’ve seen suggests that the 911 gaps were mostly in low income communities.

By this morning, a year ago, we were live updating that map. @napsgfoundation started using it to populate an app for @USCG and #Houston. The SOS calls came in faster and faster.

We onboarded more volunteers and added a team searching for pets, livestock and wildlife.

That’s when we broke Google Sheets.

Google created a special sandbox for us to use, and lent us two engineers.

We had a volunteer shallow water rescue expert in Houston, sharing our data with her team and others with some training.

They also handed data off to @USCG responders, who told us our data was exceptionally clean and useful.

When the flood waters receded, Houston Emergency Management tweeted their gratitude for our data.

We had 500+ volunteers by then, located around the world, which let us segment them into shifts for round the clock coverage.

We were tired. Ready for sleep. We had facilitated 5,000 to 7,000 individual rescues (typically with multiple people and pets).

Then came #Irma and #Maria.

We innovated a new training model using @SlackHQ, who donated support. We started a volunteer stress management program. We helped @fema use our data operationally.

By the end of 2017 we had worked with 700+ volunteers and facilitated an estimated 15,000 rescues.

We helped @fema plan, organize and run its first Hackathon.

Two years later a lot has changed.

There is now a crowdsourcing desk at @fema that activates as part of the full federal response. They are helping operationalize the work off groups like @crowdrescue.

CrowdRescue closed shop, and many of those heroes formed @CEDRdigital, a nonprofit that’s now at the forefront of federal disaster mapping.

A team of researchers from Texas A&M Galveston and the University of Delph used the online Harvey Rescue SOS map as a model to predict where flooding would take place and compared their results to conventional flood models.

Their scientific modeling has shown greater accuracy than current models and they are continuing their research to see if they can build models to provide real-time information to agencies to improve their response to flooding and keep the public safe.

You can read about this work online in the Journal of Flood Risk Management: onlinelibrary.wiley.com/doi/full/10.11…

Texas has changed its laws to support emergency response crowdsourcing and civilian volunteers.

My brother from another mother, @m_marchetti of @cs_rescue, will be directly supporting Houston Emergency Management during the coming hurricane season. (I’ll let him share details if he chooses).

For emergency managers who want to learn how to scale operations during a disaster, there’s now a toolkit: crowdsourceem.org/integrating-cr…

There are increasing opportunities for people who want to pitch in during a response - or, just as importantly, trained BEFORE a response. Here’s a recent @fema guide: fema.gov/blog/2013-08-0….

We’ve heard a lot, from hardened first responders, that “civilians aren’t helpful during disasters.”

Yet help we did.

And our groundbreaking work then will be supporting lifesaving efforts for years to come.

Cc @heathersuccio @BlueWaveCS @alt_noods @USGS_Rocks @alt_4nTrade @RogueEPAstaff @brooklynmarie @AynRandPaulRyan @AltWASONPS and so many many many others - tag yourselves and others in please!