It’s officially the spookiest time of the year.
As adults around the world set up Halloween decorations, prepare delicious snacks, and host gatherings, children are busy devising their strategies and scheming to maximize their candy haul during trick-or-treating. After all, it only comes once each year!
Given the window to collect as much candy as possible is so narrow, we wanted to give your little ones a competitive edge, powered (of course) by Minitab’s Predictive Analytics module. To demonstrate this, we imagined a scenario where data from trick-or-treaters was collected to determine which factors might be most important for maximizing candy haul.
Our Spooky Methods and Results
In this hypothetical scenario, we analyzed data from 100 trick-or-treaters. We measured the total number of pieces of candy each one collected, the costume quality (on a scale of 1-10), the time each spent trick-or-treating, the time of day, the number of friends in their group, the total mileage each one clocked, the number of houses visited, and the age of each child.
We then used Automated Machine Learning in Minitab’s Predictive Analytics module to select the best model to understand the association of the quantity of candy. In this case, TreeNet was the algorithm that made the most sense given the data we entered. Here are the results we obtained:
Most importantly, we found that the top two factors for getting the highest amount of candy were costume quality and the number of houses visited. The houses visited made sense to us, but we were surprised that the distance traveled and the time spent trick or treating were far less important than the overall costume quality.
The one partial dependence plots showed a clear upward linear trajectory based on costume quality and total number of pieces of candy, with a notable jump from costumes ranked in the “middle” numbers (4-6) to the “best” numbers (7-10). For example, costumes ranked a “6” received on average 91 pieces of candy while costumes ranked “7” received roughly 134 pieces.
Additionally, we see that the age of the trick or treater did not really seem to have a noticeable trend one way or another on the amount of candy collected. Five-year-olds tended to do the best, while six-year-olds fared the worst. Second place goes to the nine-year-olds.
R.I.P. Guesswork
Don’t you wish you had Predictive Analytics as a child? Instead of feverishly running as fast as you could with your haphazard basic pumpkin or cat costume, you could have put more time into a strategic Ghostbusters, Madonna, or even Cabbage Patch Kids costume and hit as many houses as possible.
Your kids don’t have to suffer a disappointing Halloween haul this year. And if you haven’t unlocked your Predictive Analytics module yet (conveniently located directly in Minitab Statistical Software), you can submit this quick form and select “Predictive Analytics Module.”
Happy Halloween from your friends at Minitab!
Don’t be spooked! Learn how to get the most out of your data with our training options.