I like to compare the experience of surfing Twitter to the outbound lanes of the Fort Pitt tunnel. When you get down to it, Twitter is a lot like trying to jam four lanes of 50 mph highway traffic into a two lane hole. We are trying our best to make this situation work smoothly and calmly as possible, but there is only so much anyone can do when four lanes come together into two like this. Same goes for Twitter, there’s only so much we can do with condensing our minds into 280 characters. This is why Twitter is mostly a raving lunatic asylum. It’s the 365/24/7-at-4pm-on-a-Friday-outbound-lanes-of-the-Fort-Pitt-tunnel of internet experiences.
But every so often there will be a small shining ray of light flickering on Twitter. A ray of light that gives you hope there is reason for all of this insane madness. Kind of like that guy, who looks a little bit like your Uncle Bill, waiving you in with a nice warm smile from his shimmering silver 2007 Honda Civic. Thank you, and here’s a waive back to you, Uncle Bill.
That ray of light on Twitter is Allegheny County’s twitter feed. It’s chockfull of useful information.
Especially when it comes to local flooding reports.
During many of the past flooding events in our region, Allegheny County’s twitter page was on top of things posting helpful information to the public. For example, here’s a random flooding tweet from back on June 10, 2018 from Allegheny County:
This tweet is super useful. It contains specific location information along with a date and time.
So a month or so ago, I wondered to myself: would there be a way to extract all of Allegheny County’s +17,000 tweets dating back from 2012 and filter out just the ones pertaining to flooding?
I’m an engineer (no, I don’t drive trains).
I know programming and automated scripting (a little).
I’m pretty smart (on occasion).
I think I can figure it out (or I can give up and not write this post).
Good news, I did eventually figure it out (which is why I am writing this post.)
Using automated scripts, I managed to “scrape” all of Allegheny County’s public twitter page and filter out all the ones where the word “flood” occurred somewhere in the text of the tweet.
That line above and below? That indicates where I take things to another level of nerdery. Be forewarned, beyond this point there are numbers and charts and graphs and “analyses” that experts from People Magazine say make engineers the most sexy and desirable profession on the planet. Or least, I forget which.
Graph 1 – How many flooding tweets has Allegheny County made since 2012?
The answer is 535 tweets.
Here’s all those tweets in graphical format:
Observation: Noticeable increases are shown in the summer. June and July seem to be the worst months for flooding in Allegheny County.
Graph 2 – How many individual days did Allegheny County tweet about flooding?
The answer is 149 individual days. In other words, between 2012 and August 20, 2019 (when I extracted the tweets) there were 149 calendar days when Allegheny County tweeted about flooding happening somewhere.
Here are all those dates (summed by month) in graphical form:
Observation: 2018 was the wettest year on record, but the single month with the most days with a flooding tweet was back in July 2013 with 11 total days. Over a 1/3rd of the total month!
Graph 3 – On average how many days per year does Allegheny County tweet about flooding?
The answer is 18.6 days per year.
Here are all those flooding days (by year) in graphical form:
Observation: Allegheny County will tweet about a localized flood anywhere from two weeks to upwards of a whole month every year! In 2018, over 10% of the days there was flooding happening somewhere in Allegheny County.
Graph 4 (last one, I swear) – Which months are the worst for flooding?
The answer is June and July (in a neck and neck race).
Here’s the number of flooding days between 2012 and 2019 (by month) in graphical form:
Observation: When it’s June and July in Allegheny County head for high ground when it rains hard.
BONUS ROUND!!!!
Out of its 17,000+ tweets, what is Allegheny County’s #1 most liked tweet of all-time?
At 137 likes on July 21, 2019 this tweet currently sits at #1:
Well, I wonder what is #2?
At 122 likes on May 24, 2017 this tweet currently sits at #2:
The Pittsburgh Penguins are losing to regional flooding problems on Twitter.
What a dumpster fire.
Oh wait…
That is all for now…
(BTW, I plan to be using this Twitter data to make some neat maps in the near future. Stay tuned.)
The Pittsburgh Urban Flooding Journal 
Leave a comment