Three years ago, I started work on a PizzyLabs project for sending real-time weather bulletins over Twitter’s Direct Message feature for users based on their desired locations. That was the basis for the @NOAAalerts twitter account.
Many things happened shortly after I got started with it; I figured out it wasn’t feasible to check every NOAA RSS feed every minute for instant updates, I didn’t have as much of an understanding of User Interface components as I thought I did, the functionality of the CRON jobs that would be required was very difficult to implement, and then my most dreaded personal issue: I solved all the problems in my head, and it took away all the fun of actually doing the work (aside from the unfeasible parts).
Here we are in 2014 though. I’ve had a job developing user interfaces for the past 3 years, and the things I create at work are celebrated by all the important people above my position, and our customers. I’ve also taken the SKYWARN StormSpotter training for 2 years running. And I had another idea that’s much more feasible:
StormSpotter Activation Forecasts
The one thing that bothers me the most about being a StormSpotter is the fact that I never actually know when the NWS will activate us. I have to manually check the Hazardous Weather Outlook for the day and see what the forecast is for activation.
So why not write a script to do that for me? I looked all over the internet for an API to call so that I didn’t have to scrape their pages, but I couldn’t find anything. I did find a lot of other cool stuff though that I would like to integrate into other ideas.
Well, it’s only been a week, but so far I’ve gotten a rudimentary UI in place, with Twitter Sign In, and NWS Office city/state selection. I have a script that can read the HWO and parse it for counties affected and Spotter Activation Statements. And I have the Twitter Application code set up to send the appropriate DMs to users once their settings are chosen.
Things were much easier this time around than last, and the one thing that has been the strangest is I haven’t given up on the project just because I understand how to do all the aspects of it. It’s definitely a challenge to get back into MYSQL after spending 3 years working with MongoDB, and it’s also a challenge to write my own PDO class for handling the queries and a template class for managing the UI (I haven’t done anything with Frameworks before, so that was off the table for now).
I have made the code open source over on Github, since there really isn’t any reason not to…I’m still new to the Open Source code thing, so I’m probably using Github wrong, or not sourcing the right information. Leave a comment below, or over there if I’m missing something, and I’ll get it fixed.
I’m not ready to announce the service yet, since the CRON jobs aren’t in place yet to send the forecasts. However, the work in progress is accessible from the PizzyLabs page if you’re interested in giving what *is* there a try. I’ll have a full write-up once it’s released.