Predicting the future with bubble

Hey bubblers

I’ve designed a form for a friend which is super simple.

It’s a 15 question form which he uses for his work.
He has to conduct this 15 question checklist every day.
I made this form for him because it’s just a lot quicker than doing it on paper.

There is one question in particular on the form which says.

“Is the pressure gauge reading below 27 psi”

Now if it is a yes then its a FAIL so, I save that as a 0 in the data base.

We’ve discuss that we would like to show him everyday a percentage of failure.

For example. Is it possible to show him a simple line of text on the form which says that " There is 69% chance that the reading will be below 27 today".

Is there a way to predict this? And how should I approach this?

Any advice would be appreciated.

Hi there, @timgarrett111… it’s certainly possible to show a simple line of text like you mentioned, but the question is what is the predication based on? Is it just a simple average of all of the pressure gauge readings? So, if you have 10 days of readings and 3 of them were below 27 psi, would the prediction on day 11 be that there is a 30% chance of the reading being below 27 on that day? Or does the prediction have to take into consideration some of the answers to the other 14 questions on the form on any given day?


Sure. There’s a better way to do the logic in bubble, but here’s a general idea.

This bold workflow will display 1 day Failure rate.

Do a search for (PRESSURE_READING) Where value = 0 AND date created <1d ago:count
Do a search for (PRESSURE_READING) Where date created < 1d ago:count

10 failures / 50 total = 0.2 = 20% failure rate.

(Edit: I know i’m doing the math for the Past - - - But we use the past to predict the future :slight_smile: )

Hi @mikeloc @jobs

Thank you so much for your reply!

Wow! Ok I completely understand that. This is actually really powerful.

@jobs So how would it look from a calculation perspective if I wanted to workout a percentage/probability that there will be a failure over the next x amount of days. (Let’s just say 30 days)

So for example.

  1. If this checklist is taken in the next (30) days there is a (%) chance that it is going to fail.

Is there a way to predict that?

I really do appreciate your help here.

Kind regards.

I think that’s more of a data issue more than a bubble issue.

Predicting a failure is tough. If you’re looking for the most accurate way you’d want to gather as much data as you can from the devices ( usage / date / weather / device serial# / whatever ) and put it through a Machine learning identifier to detect probability of failure.

Or if you just want it to look cool you could check the overall devices failure rate during past month = 5% so chances of device failure for this month is 5%.

Yes. I don’t want to go down the road of measuring hardware or anything that far to in depth.

Keeping it simple and looking cool while providing some value is the thing we need.

Thanks @jobs your a champ! Your input has been really helpful for me. I really appreciate it :pray::pray::pray::pray::pray::pray::pray:

1 Like

This topic was automatically closed after 70 days. New replies are no longer allowed.