Hayo, I'm moving this to it's own thread as I feel it's more suited in it's own thing than taking up replies in another thread. So, I am simply copy-pasting what I put over there onto this thread. (excluding some now unnecessary information)

While my intention is not to directly advertise a twitter account, I developed a twitter bot that automatically posts a few interesting things relating to Kings Island (hence some KIC members might be interested)

1. Every morning, the bot will post the operating hours of the day and when the park will open. If the park is not open, it tells you when the next operating day is and when the park will open.

2. Whenever the park opens or closes, it tells you either how long the park is open for and when the next opening date will be.  

3. During operating hours, the bot will post wait time updates once every hour, at the half past mark. (10:30AM, 11:30AM, 12:30PM, etc) until close. 

4. Alongside these operating hours, the bot will post snips of the official Kings Island webcams (found here: https://www.visitkingsisland.com/live-video)

The wait times run off of the same data other wait-time sites use - like queue-times, thrill-data, or isitpacked. The images are still-images fetched from the video feeds directly and sent as media. 

The reason I developed this bot is solely so myself and others can see current wait-times in a quick way, how busy the park is, and other info right from twitter whenever I'm scrolling through it. Hope you'll find it kind of neat like I do, as you may even be able to see yourself if you time it right. If you are interested in seeing these automatic updates on your timeline, feel free to give the bot a follow on the handle KIWaitTimeBot.

My long-term goal is to have it also post end-day statistics in a visualized format to get an idea of how busy the park was for any given day, similar to how thrill-data and queue-times do it.

The web-cam snip part is brand new as of yesterday and for an unknown error, the 6:30PM webcam snip did not function properly. Am debugging!

You developed this? Very cool!!!

Edit: I couldn’t find it by searching. Could you link please?

On 7/2/2022 at 5:09 PM, BoddaH1994 said:

You developed this? Very cool!!!

Edit: I couldn’t find it by searching. Could you link please?

Yes, https://twitter.com/KIWaitTimeBot 

It runs on a Node VPS using Twitters v1.1 API. 

This bot has been extremely helpful. Use it to gauge whether or not I want to go to the park when I'm free.

Awesome bot! Kinda sad I didn't think of something like this, looks like a fun project!

Amazing project!

 

Wow, you can really see how much Beast has been down!!!

TIL:

1. The webcams are back on, even though the page still says they're not.

2. One of the lights on Orion's helix appears to be out at the moment.

Due to the unfortunate removal of webcam screenshots, I have developed a secondary replacement of sorts. 
Though, nothing can really replace the real-time screenshots that were provided hourly. :salute:

See full update tweet thread here: 

 

I've season FL+. If it helps, I'd be happy to compare measured against your predicted FL(+) queue times.

@Coasterteam - since it seems to be a cool hobby you are working on and developing algorithms, is there a way trusted members here could provide input as @NegativeGs suggested for FL+?  

Maybe with a timer and from when someone enters the line until ready to get on a ride?  While that site does allow users to put in times, some are jokes like the almost daily one at CP of someone saying they waited 5 minutes for dragster.

And then you could develop an algorithm that takes trusted times into account?

Of course it's accuracy could be off because there are so many variables that come into play, but more data input into can probably start a predictive model with a little more accuracy.

On 7/9/2022 at 4:42 PM, disco2000 said:

@Coasterteam - since it seems to be a cool hobby you are working on and developing algorithms, is there a way trusted members here could provide input as @NegativeGs suggested for FL+?  

Maybe with a timer and from when someone enters the line until ready to get on a ride?  While that site does allow users to put in times, some are jokes like the almost daily one at CP of someone saying they waited 5 minutes for dragster.

And then you could develop an algorithm that takes trusted times into account?

Of course it's accuracy could be off because there are so many variables that come into play, but more data input into can probably start a predictive model with a little more accuracy.

That’s a good idea. I’m open to suggestions on how to improve the algorithm. 

While I can’t exactly predict it to the second, like how queue-times tries to predict any specific days crowd level but it’s actual level can vary greatly.

What I wanna add in next is variable capacity numbers and changing guest interest depending on the time of day. An example being guest interest for Beast increases at night. 

 

EDIT: This was added in. Dynamic guest interest & dynamic FL capacity included. Estimating FL amounts based on the queue-times predictions + my own predictions to get solid numbers. It's rather magical.