New Feature idea... feedback wanted

Hi, I’m Steve Bloom, one of the statisticians that work at TouringPlan. The Crowd Calendar predictions are long range predictions. We know that weather has a big impact on crowds and crowd patterns. We could publish a short term 10-day Crowd Calendar using weather information in the modeling. The predictions would be updated each day. Just like a weather forecast, predictions would be more accurate as a day approaches.

Would changing crowd levels be too confusing?

Would this be helpful while you are on your vacation?

7 Likes

I would find it very helpful to have weather information published on the site! Thank you for all your hard work!

1 Like

I think it would be helpful. That is if I could access the crowd calendar at all-it is all gobbly gook. Would it be part of the regular crowd calendar or an additional one? if part of the same maybe make it a different color to differentiate ?

1 Like

Sorry about the gobbly gook. For some user with Chrome browsers they are getting the unformatted page. The current workaround is to use http://touringplans.com/walt-disney-world/crowd-calendar2.

We are still trying to figure out how we would display short-term and long-term predictions. One idea is to only show the short-term predictions in the Lines app on the 10-day screen.

4 Likes

Thanks you guy ROCK.

I would love this. :slight_smile: A very, very awesome idea.

Sounds like a cool idea!

I thought the crowd levels were updated with the latest, best predictions. I understood that the numbers (crowd levels, wait time predictions, etc) changed with conditions up to and including the day of touring. Since you guys aren’t including weather as a factor in your predictions, that suggests to me that your forecasts could and should be better.

As an aside, I’d rather see a mean with a standard deviation which would no doubt be closer to reality. Or, per your weather analogy, you could say something like ‘there’s a 70% chance of a crowd level 8’ or whatever makes sense.

More accurate, constantly updated crowd forecasts means better wait time predictions which means more efficient touring. Right?

Hmm… Although up to date information should be made available, it won’t change my plans. Given how far in advance I need to make all of the plans, I really rely on the crowd predictions much further out and am very unlikely to change my plans based on updated information at the time of my vacation. I rarely do park hopper option and I book dining and FPP, so I am pretty locked in.

But, wouldn’t it change your plans on the day of touring as you reoptomize though the day? Suppose rain was likely at a certain time, and TP knows that the park with be less crowded and therefore the newly optimized plan should allow for reduced wait times? Or, perhaps the plan could send us to indoor attractions that don’t close if- say- lightening were predicted or users input that several outdoor attractions are closed (likely because of weather). Etc.

1 Like

First, you are awesome! I <3 Touring Plans.

On to the question at hand, I think whatever data you can incorporate to make the crowd calendar as accurate as possible is great. I anticipate that those numbers will regularly change, so this model fits my expected user experience anyway.

While having weather taken into account is good, I agree with @Iheartepcot that info may not be actionable. My ultimate decision of what park on what day has to be made 60+ days in advance, so I can get FF+, ADRs, and all that jazz set up. Having crowd calendars update based on 10 day forecast would only impact if/how we park hop later in the day.

Thank you for all you do.

1 Like

As for how to present it, I wouldn’t keep the revised short-term predictions just in an obscure location in the App. It might make sense to have parallel columns in the crowd tracker with a different (smaller?) colour/font that notes “Adjusted for weather” so that users have access to both old predictions (which I used when I made my TPs) and new prediction (so I can decide if there’s enough of a difference for me to change my plans in the days leading up to a trip). Also, would this translate into the ability to optimize TP’s for adjusted-for-weather predictions, too?

2 Likes

As a sorta local I make last minute plans/changes a lot and this would be very helpful. Look forward to it!

More information is always better for me, so I’m all for sharing this.

[quote=“SteveBloom, post:4, topic:15219”]
We are still trying to figure out how we would display short-term and long-term predictions. One idea is to only show the short-term predictions in the Lines app on the 10-day screen.[/quote]

Whatever you do, I’d like some easy way to know if predictions changed and by how much, maybe showing a number in green if predicted crowd level drops or red if it goes up. Some indication of the magnitude of the change would also be useful.

2 Likes

Yes, the re-optimizing feature, if updated to account for weather - or ride closures that happen for other reasons without warning - would be super helpful. There have been many afternoon showers where I am scrambling to figure out which attraction to go to when I otherwise have an outdoor option scheduled. It’s more that changes to overall crowd level for a park would not change my touring schedule. I love to re-optimize throughout the day, particularly when things take longer or shorter than expected, usually due to us procrastinating or getting distracted from our plan.

1 Like

I think the short-term updated idea is great.

I have to admit I am struggling with this one. I expect I would like the information, it may help with my PH plans, but I do not know if it will actually change how I plan. FPP has me pretty much locked in at 60 days. It may have me bump a FP to a later time if I can, but I do that now. For me, I think of AK. After a recent experience, if a big rain storm was coming through, I would know that park projection would be very low but honestly, I hope to never find myself in the park during torrential rainstorm. So, I guess my answer would be , yeah, I would like the information, but I do not know what profile of WDW visitor it would help the most and would it fit on my dashboard?

1 Like

Sorry, I do not like this idea at all. 10 day weather forecasts are useless, as anything beyond 3 days is about as accurate as a Ouija board. I definitely would not want to see a 10 day CC that had CLs adjusted based on the weather forecast.

Have you done any analysis of this, comparing “base” predictions and “weather adjusted” predictions versus the actual CLs? If there was any statistical support for this I might be more interested.

1 Like

Point well taken. I agree that 10 days out weather can change a lot. It might be better to just update the next 3 or 5 days out. In either case, we would update all days in the near future every day.

Every day we look the performance of our crowd level and attraction wait time predictions. (Our crowd levels are based on attraction wait times.) When the predictions are off, most of the time we can attribute the discrepancy due to weather. If Monday and Tuesday have thunderstorms, and Wendesday is sunny, the crowds will be higher than expected.

2 Likes

Currently we do make adjustments of the current day’s attraction predictions. The adjustment are based on the observed data collected.

Adding weather to the mix would help for those afternoon storms that role in and shut down some attractions.

We are always looking at ways to improve our predictions. Even if we don’t calculate a short-term crowd level, we will do everything we can do to improve the attraction predictions that are used in the Lines app, and the Optimizer.

2 Likes