Altair RISE

A program to recognize and reward our most engaged community members

Nominate Yourself Now!

Rapidminer to predict for smart parking

Hi all, I'm very very new to rapidminer (literally 2 days old user) and would like to explore more about rapidminer. I'm trying to do a predictive analytic for smart parking. I would like to have a simulating model for predicting the occupancy of parking garage and predict turnover for the parking garage. Is it possible done by rapidminer? I was using R language before this, however, it spent too much time on writing codes.

Much appreaciate!

Thanks!

Find more posts tagged with

AI Studio

Accepted answers

All comments

MartinLiebig

Hi,

this is possible. THe question is if you have the data available.

~Martin

siew735

Hey Martin! Thanks for the feedback, I really appreciate it a lot.

A quick question, what is the min. duration of data if I want to do predictive analysis? I understand that big data do need data for like the past 5, 10,20 years. However, is it possible to done predictive analytics by using minimum data such as 7 days, 2 weeks or 1 month?

Thanks a lot, once again!

MartinLiebig

Hi,

there is no minimum duration. It's rather a minimum number of examples. It would be good to have a few thousands of examples for both classes representing all "states" of the problem.

~Martin

siew735

Hi Martin. I see, haha. Cause we currently do not have much data, parking occupancy data for approx. 7 days only.
Sorry for bothering but how do i know what variables I should have in the data sets? What are the parameters needed to do its predictive analytics? haha
Thank you Thank you for all the replies

MartinLiebig

Hi,

short answer: I have no idea.

long answer: The point of advanced analytics is that you do not know before hand what will work best. The only thing you can do is think like this:
In Predictive Analytics you always extract general rules out of your data. A general rule might be "if the avg turn over frequency is 1 hour and the garage is currently empty for 1 hour, it is likely to be taken again".
So the only thing you can do is think about this (in a group on a white board with a coffee) and think "What could the rules be?" and then try it out. Feature selection helps to remove features. Generating new one is usually data scientists/domain experts work

~Martin

JEdward

I'd also suggest adding a pizza to that meeting, it always helps to write out the ideas when eating cheese.

Also, I'd say don't just limit your dataset to the garage itself, are there any environmental variables that you should take into account when collecting the data?
For example, if the parking garage is near a shopping centre is the centre busy at certain times of day & would they have any data available which you might be able to check correlation to your parking space occupancy?