The quiz question
You have an ice-cream shop. You noticed a relationship between the number of cones you order and the number of ice-creams you sell. Is this a suitable situation for regression analysis?
has answer as “NO”
Could you clarify why it is so?.
The reason I think it is yes is..
As the sales increase the cones demand increases so the owner order more cones to serve ice cream.
So if I have the data of cones ordered and the sales figure, based on regression analysis the future sales can be predicted based on the cones the owner ordered.
Hi Amar,
thanks for reaching out!
Unfortunately, the number of cones is not a good predictor. It has no predictive power, but instead is dependent on the output variable. That means that the sales of ice cream determine how many cones are ordered. But not the other way around.
Consider the following:
the owner of an ice cream shop buys a large amount of ice cream cones. But if it’s in the middle of winter and its cold outside, so no one comes to buy any ice cream. So, the number of cones bought is not a good predictor for sales.
Best,
Eli
In addition to that, another big issue is that these are interrelated. Each ice-cream (usually) needs a cone. You are not predicting one with the other. It is the same as having a pair of socks. You don’t predict the number of left socks by the number of right socks.
In the same way, ice-creams and cones are one and the same thing – they go together!
Hope this helps!
Iliya