Sandy Search for the Call for Code Hackathon
Sandy Search (was Emergency Services - Search and Reporting) Mobile and Web App
Entry for The Global Call for Code Hackathon, 28 September 2018
I was in Super Storm Sandy in 2012 in Connecticut, and was without power for 9 days. During the disaster, we had a very hard time discovering what services (gasoline, food, water, ice, power) were available and when and where. And if you have only a half a tank of gas you don't want to drive around looking for more, without knowing where to go.
The idea for this service is to use the power of crowdsourcing (think "Waze for Services, after a disaster"). It currently allows anyone (without a login) to search and report on services available in their area.
All users who do not login will be anonymous and not logged.
Verified users (pre-registered, think Red Cross, National Guard, and business owners/managers) can leave "verified" status updates, at no cost.
I think the service could be used as a branding expense for emergency services. The majority of the expense would be the manual verification of "verified" users, and minimal hosting of the back-end service.
How It Works
So this service uses geolocation and a standard crowdsourcing format to collect information from people in the field. And share it with anyone who needs it.
One of the driving forces in the design was simplicity. "2 clicks to get to the answer", exactly like the Palo Alto Police Chief said in this video
Nothing fancy but it will work, much like the way Zello was used after Hurricane Harvey in Houston: https://wgntv.com/2017/08/29/civilians-and-cajun-navy-bring-their-own-boats-to-rescue-harvey-victims/
How It Was Built
This mobile application was built with the sweet Ionic 3 Framework (Angular5/TypeScript/CSS/HTML5) and uses the Google FireBase services for authentication, realtime database and cloud hosting of the web app (currently on a free spark plan).
Currently generating Android app and will provide iOS app from same code base.
Maps and geocoding is pulled from Google APIs. I plan to do reverse geocoding from Google too.
Admin user interface is all Node and Angular. And hosted in a Docker container (stored in the IBM Cloud Container Registry) in Kubernetes cluster on the IBM Cloud Kubernetes Service. Need to add reports and maps to admin user interface.
Want to use IBM Watson services to post-process all user request data and create heatmaps for the optimal locations for services and food distribution.
All built with a Vim editor and a cmd window (and git).
What I Learned from this Hackathon
- Angular 4 and Ionic 3 and typescript (I had build the previous prototype in Angular.JS and Ionic 1)
- Firebase Database with Observables
- GeoFire for geolocation
- Angular 2 and NodeJs
- Dockerizing an Angular 2 / NodeJs application
- Autobuilding a Docker image from a GitHub checkin with Docker Cloud
- Storing and updating Docker images in the IBM Cloud Container Registry
- Running Apps in Kubernetes in the IBM Cloud Kubernetes Service (on the free tier) with NodePort networking
- Fixing vulnerabilities discovered by the IBM Cloud Container Registry Vulnerability Advisor
- Upgraded running Docker containers with new images
- Upgraded running Docker containers in Kubernetes in the IBM Cloud Kubernetes Service with new images
Next Steps for SandySearch
- Add reports
- Watson ML to predict where services will be needed based on past experience
SandySearch Submission Video - First attempt - https://youtu.be/83FHRSM3CQM
SandySearch Submission Video - Second attempt - https://youtu.be/Vj61wD86HyA
SandySearch Submission Video - Third attempt - https://youtu.be/kELeIpW1EdA
SandySearch Admin User Interface Demo - https://youtu.be/hGE8ubJQGMg
Sandy Search Mobile/Web app
Live web app version of SandyService can be found here: https://plu.sh/sandys
Live SandyService web app hosted on IBM Cloud Kubernetes Service can be found here: http://220.127.116.11:32705
Note: to run this in a chrome browser (and have geolocation work on an insecure URL) use the following:
And the Android APK is available on GitHub here: https://github.com/mkobar/sandysearch/releases/tag/0.0.7
And on Google Play here: https://play.google.com/store/apps/details?id=io.ionic.sandysearch
SandySearch is now on the Apple App Store too: https://itunes.apple.com/us/app/sandysearch/id1438425201?ls=1&mt=8
The original GitHub repository for the Ionic v1 web app is here: https://github.com/sandysearch/essr-aoc
Sandy Search Admin User Interface
Live web app version of SandyServiceAdmin (v0.0.4) can be found here: https://plu.sh/sandyadmin2
Live web app (v0.0.2) running on IBM Cloud Kubernetes Service can be found here: http://18.104.22.168:30471
Live web app (v0.0.3) running on IBM Cloud Kubernetes Service can be found here: http://22.214.171.124:30433
Live web app (v0.0.4) running on IBM Cloud Kubernetes Service can be found here:
Sandy Search Docker image on Docker Cloud: mkobar/sandy-search [private]
Sandy Search Docker image on IBM Cloud Container Registry: registry.ng.bluemix.net/call4code/sandysearch
Sandy Search Admin Docker image on Docker Cloud: mkobar/sandysearchadmin
Sandy Search Admin Docker image on IBM Cloud Container Registry: registry.ng.bluemix.net/call4code/sandysearchadmin
Things left to do:
store service location data in firebase
add a map view (with Google Maps or OpenStreetMaps)
add the account request function (and screen) - and force to validate with email?
add more logging
add in the real distance calculations (from street address). There are several free services for this but they may be rate limited.
add dupe reporting option
add initial tutorial slides
fix km to miles
add an admin interface for entry and user management
- [ ] rate limit votes and disputed to 1 an hour with local storage
- [ ] add spinner for list and map
- [ ] change markers by serviceType
iOS build and App Store upload - just need my Mac and a day
Dockerizing an Ionic 3 app
Google Play Store upload
host admin interface on GitHub
host admin user interface on IBM Cloud (Docker and Kubernetes)
- [ ] add reports (timeline maps by createDate, updateDate and votes)
- [ ] export usage data in csv format
- [ ] add Watson ML for heatmap generation
- [ ] process usage with Watson to predict optimal locations for future services
need a video of both app usage and reporting
need better logo ;^}
Please do contact me directly if you can use or add to this project.
Thanks to Dick Tator of Fayetteville, NC, USA for the original logo.
Copyright @ 2018 RKOSecurity