In 2017, Walt Hickey from FiveThirthyEight set out to find the Halloween candy most people prefer. To figure it out, he set up an experiment in which online voters could select their favorite candy from randomly generated matchups (for example, Twix or Snickers?).

After compiling about 269,000 votes from 8,371 different IP addresses, Walt calculated the win percentage for each type of candy and included several attribute (for example, does it contain chocolate?).

Recommended Analysis

• Which candies have the highest win percentage? What do they have in common?

• Are more expensive candies more popular? What about sugary candies?

• Can you segment the candy into specific groups?

• If you had to pick 3 candies to give out in Halloween, which would they be?