Hurricane Sandy: Ushahidi + IR build
This past weekend, OpenIR participated in the #HurricaneHackers #Sandy Crisis Camp Hackathon to aid with the hurricane crisis response. Our output: a New York City data viewer, with IR layers and geotagged tweets, built on Ushahidi 2.5. This Ushahidi build is more of a research simulation; it was made too late to be used during Sandy, although it is currently collecting some useful geotweets about recovery needs. As the tweets start to populate the map more densely, we will study how they cross-reference land features highlighted by the IR data.
OpenIR does now know how to quickly compile Landsat IR combinations, build them into an Ushahidi deployment, and add geotweets– things that we can now do at the beginning of future disasters. When we get back to developing vulnerability maps like that in our Jakarta Ushahidi deployment, that will also be a great research tool.
Part I: Setting up Ushahidi + IR layers
First, we set up an Ushahidi build with the newly released 2.6 platform. However, our internal OpenIR plugin, which worked so great with Ushahidi 2.5, didn’t work in v.2.6. We’re starting to work on fixing this, but in the meantime, we installed and set up our infrared layers (TMS) in v.2.5. Our messy plugin, for internal use only, is on Github. There is no UI for uploading and replacing TMS layers, this is modified in the plugin code.
From the admin dashboard, we also added news streams (NYTimes City Room and NY/Region, plus WNYC’s Brian Lehrer show), changed the theme, un-clustered reports, and made a few other tweaks.
Part II: Adding Categories and Manual Reports
Using Kate Starbird’s great Tweak the Tweet for Sandy as a model, we set up report categories and started manually reporting. We didn’t get that far in the couple of hours spent hacking together this data viewer, and it seems that the geotweets (see below) are providing much extensive information that we could manually. For this project, and perhaps for future OpenIR Ushahidi builds, manual reporting may best be used to make observations about how the IR layers are useful/relevant/helpful for the deployment.
Part III: Importing Geotagged Tweets
We integrated geotagged tweets with the hashtags #Sandy, #Hurricanesandy, and #Frankenstorm, which are represented by the red dots, into the map.
To automatically turn a geotagged tweet into a report in the data viewer, we first had to choose which tweets were relevant. We picked the hashtags #Sandy, #Hurricanesandy, and #Frankenstorm in the Settings tab of the admin’s Dashboard.
Then, we went to Manage>Actions and created a Trigger.
Now geotagged tweets with Sandy related hashtags will automatically be added to the NYC data viewer.
To find out more about Hurricane Hackers, go to their linklist.