My Halloween Candy In Graphs

Halloween usually means candy.  Less commonly, Halloween means making geeky graphs on the distribution of candy you give out to Trick-or-Treaters.

This is the first Halloween at my new house, so we didn’t know what to expect.  So, why not graph out everything?

The Stats

  • 419 Treats, 379 Treats taken by Trick-or-Treaters
  • 189 Trick-or-Treaters
  • 2 Hours of Trick-or-Treating
  • 2.01 Treats taken per Trick-or-Treater

Total Candy

Yes, I counted all the candy.  Geekier still, I graphed the starting percentages and ending percentages.

What does it all mean?  Are sugary candies less popular in my neighborhood than straight-up chocolate?


We had 189 Trick-or-Treaters spread over two hours.  Notice the inverse relationship that the more kids that visited within that 15-minute block, the less candy they took and vice versa.  We told each Trick-or-Treater to take two, and some listened better than others to that.  We also think that in the busier times it was a bit more crowded, so perhaps some kids wanted to just take one and move onto the next house.


In prep for Halloween, we bought lots (and lots) of candy.  Nineteen types, in fact.  Some bags were single types, some bags were multi-packs.  You’ll commonly find multi-packs in stores and it’s an easy way to get a diverse mix, but we counted up the breakouts of the different types that we bought.

Favorites and Non-Favorites

Most of the treats were taken, but there were some left and some that went faster than others.  How do you graph this?  We wanted a better indicator, so instead of straight numbers, we calculated the percentage of that treat to the total amount remaining.  We were able to get a count of the remaining treats about 45 minutes in, and then made time notations of when we ran out of that candy.

You saw above what was left over, but how do we graph the change over time?

Well, we put all 19 types together on a stacked bar chart: 

You can kind of see the trend, but it’s not exactly clear.  With repeating colors, it’s hard to tell what is what.  Which red is the one on the bottom?  Is that Milky Way?  Almond Joy?

What about a line chart?

Slightly improved, but not great. Yes, you can see the outliers on the end of the chart, but 19 candies is apparently too much for a line chart.


Grouping can make it better.  Placing the different treats into groups helps you more easily see the trends that exist.

  • Chocolate – 3 Musketeers, Almond Joy, Almond Joy Pieces, Butterfinger, KitKat, M&M’s – Plain
  • Crispy – Goldfish, Kettle Corn, Rice Krispies Treats
  • Nuts – M&M’s – Peanut, M&M’s – Peanut Butter, Reese’s, Reese’s Pieces, Snickers
  • Sugar – Caramel Apple Pops, Skittles, Starbursts

Sure, you could group them in different ways but I thought that these were logical splits based on the types that they were.  Brand could be another way, but my assumption is that most kids wouldn’t have a brand-based loyalty.

Here’s that stacked bar chart again, but by group.

A bit better in highlighting trends over time, perhaps?  You can see the crispy treats disappeared around 8:00pm.  The proportion of chocolate and nuts gets lower and sugar gets higher.

Here’s the line chart again by group.

Overall, I think the line chart highlights even better the increase of the sugary-based treats in the pool of the remaining, plus it allows you to see the slow drop-off of the other types of treats.

Reflecting on why that is, we have a few guesses:

  • We had buckets where we separated the chocolates and nuts away from the sugar, to give kids that may have had a nut allergy an easy choice.
  • Some of the chocolates were in a cauldron-shaped bucket that had a larger opening that a trick-or-treater could see what they were choosing easier.  The sugar treats and the M&Ms were in two jack-o-lantern shaped buckets that were a bit more difficult to see inside.  Perhaps this is a case of imperfect information?  If the kids could see the buckets better, would they have made more informed choices?
  • The Starbursts were the most leftover, but they were smaller than the other treats, so did the treats fall to the bottom of the buckets?

I’ve already turned this into the geekiest Halloween that I’ve ever had, so these questions may have to wait until next year to be answered.  Now if you excuse me, I have some leftover candy to eat.

9 Replies to “My Halloween Candy In Graphs”

  1. If you were a woman, I’d kiss you. Right on the mouth. This is a wonderful post that more people should do. And the findings were pretty neat. Who would’ve though crispy would get demolished faster than sugary? Not the folks at Starburst, that’s for sure.

  2. I think that the goldfish got a bum deal considering the lack of inventory. also despite crispy in the name rice krispies treat are more sugar in my mind. also i would like a carmel apple pop to be collected this weekend.

  3. Some assumtions i think you should revisit for next year: 1. I think kids can be influanced by brand loyalty as kids are directly marketed to, but in this case all you candy is brand name, as apposed to some no name candy, or less popular named candies ( tho almond joy is not common in ottawa canada) so i dont think that is a concern here 2. Kids are influenced by the size of the packaging, the bigger threats like crispies seem like a better pick to a kid because theyre bigger, regarless of actual amount of candy ( grams or oz). 3. I agree that containers should all be equal as you mentioned in you analysis. 4. In future years, i think to get cleaner results you should pick one type of candy for each category… Probably your most popular from this year, or is that going too far??? 🙂

    I really loved this graph, and im glad you did it. I look forward to future graphs. I’m amazed at the volume of trick or treaters u have, it’s a good sample size.

Leave a Reply

Your email address will not be published. Required fields are marked *