Maven Halloween Challenge (October 2024)

The Task…

Using online votes ranking 85 types of candy, the task is to find the 3 treats I will give out on Halloween to guarantee that trick-or-treaters of all tastes find something they'll love and present the data to back up my decision.

The Approach…

To find the 3 treats that I will give out on Halloween, I used a 3-step approach.

1. Divided the data into 3 exclusive categories -
   1. Allergy-safe treats (Spooky Safe Sweets) - includes candies that do NOT contain chocolate, caramel, nuts, nougat, and crisped rice and wafer;
   2. Chocolate-rich treats (Dark Magic Munch) - includes candies that contain chocolate and/or one or more of the other flavors: caramel, nuts, nougat, crisped rice and wafer;
   3. No-chocolate-no-sugar treats (Wickedly Low Sugar) - includes treats that do NOT contain chocolate. To ensure that the chosen treat has the lowest amount of sugar, I took additional steps in the following phases.
2. Normalized the data for win percentage, sugar percentile, and price percentile in each category and assigned different direction to the 3 metrics in each category:
   1. Allergy-safe treats (Spooky Safe Sweets)
      1. Win % - higher is better;
      2. Sugar percentile - higher is better (these candies lack most of the flavors kids like, so I thought they would appreciate the extra sweetness);
      3. Price percentile - lower is better;
   2. Chocolate-rich treats (Dark Magic Munch)
      1. Win % - higher is better;
      2. Sugar percentile - higher is better;
      3. Price percentile - lower is better;
      4. Number of different flavors - higher is better; I introduced this metric to be able to give a little advantage to the chocolate-rich treats that offer the most flavors in one bite!
   3. No-chocolate-no-sugar treats (Wickedly Low Sugar)
      1. Win % - higher is better;
      2. Sugar percentile - lower is better;
      3. Price percentile - lower is better;
3. Calculated the total score in each category by adding the weighted scores of the metrics in each category and sorted them from the highest to the lowest score. Please note that I assigned different weights to the different metrics and they are as follows:
   1. Allergy-safe treats (Spooky Safe Sweets)
      1. Win % - 0.7
      2. Sugar percentile - 0.2
      3. Price - 0.1

In this category, unlike the other categories, I assigned the highest weight to the win % because I thought that given the restrictions and limited selection in terms of flavors, it would be best to pick the one that most would like!

2. Chocolate-rich treats (Dark Magic Munch)
3. Win % - 0.5
4. Sugar percentile - 0.2
5. Price - 0.1
6. Nbr of flavors - 0.2
7. No-chocolate-no-sugar treats (Wickedly Low Sugar)
   1. Win % - 0.2
   2. Sugar percentile - 0.7
   3. Price - 0.1

Since this is going to be the parents’ choice - especially to those who do not allow their kids to have chocolate and sweets, I assigned the highest weight to the ‘sugar percentile’.

The Design…

I presented the results in the form of a riddle game and created the visuals so that they fit my theme: "Candy Cauldron Riddle". I used ChatGPT to generate the riddles (and everything that involves creative writing!) based on the information I inputted after calculating the total scores. The user will need to solve a riddle for each category to find out what are my 3 candy picks that I will give out on Halloween:

1. Allergy-safe treats (Spooky Safe Sweets)/"The Fruity Fright": Skittles Original

2. Chocolate-rich treats (Dark Magic Munch)/"The Nutty Dilemma": Snickers

3. No-chocolate-no-sugar treats (Wickedly Low Sugar)/"The Caramel Conundrum": Werther's Original Caramel

To visualize the data for the top 3 candidates in each category, I used a modified version of bubble chart and check boxes.

1. Bubble chart - I used the bubble chart to be able to plot the normalized scores of the metrics in the form of ‘bubbles’ and maintain the proportion depending on the value. Then I stripped all the other elements of the charts and used only the bubbles that I positioned over the cauldrons.
2. Check boxes - to give further clues about the candies and add a little more flavor to the 'Candy Cauldron' theme of the report - instead of using ‘actual’ check boxes, I used little ‘fires’ which signifies whether or not the candy has this flavor;

Lastly, to go along with the theme I picked blue (‘nighty’) and orange (‘pumpkin’, or ‘Halloween’) colors and basic Halloween-themed icons that I modified to get the desired effect.