__STYLES__

Candy Crush Insights - Halloween Favorites

Tools used in this project
Candy Crush Insights - Halloween Favorites

About this project

Project Overview:

In this project, the goal was to identify the top three Halloween candies using three key analyses: Correlation Analysis, Sentiment Analysis, and Multi-Criteria Decision Analysis (MCDA). The project focused on analyzing candy attributes (e.g., chocolate, sugar, caramel) to assess their influence on consumer preferences and determine the best Halloween treats.

1. Correlation Analysis

This analysis evaluated the relationship between various candy attributes (like chocolate, peanut/almond, and fruity) and the popularity of each candy, represented by the winpercent (the percentage of consumer votes a candy received). Using the Pearson correlation coefficient, attributes with strong positive correlations (e.g., chocolate, peanut/almond) indicated higher consumer preference. Candies with higher correlations to popular attributes were ranked higher in this analysis.

2. Sentiment Analysis

In this analysis, the winpercent was used as a direct reflection of consumer sentiment, representing the percentage of voters who favored each candy. The candies were ranked purely based on their winpercent, making this analysis a measure of overall popularity. Candies with higher winpercent were viewed as the most positively received, showcasing the direct preferences of consumers.

3. Multi-Criteria Decision Analysis (MCDA)

The MCDA approach took a more comprehensive view, evaluating candies based on multiple factors, including sugar content, price, and key candy attributes (like chocolate and caramel). Each factor was weighted based on its importance, allowing for a balanced evaluation of each candy’s overall performance. This analysis considered not just consumer popularity but also practical factors like sugar and price, making it a more holistic assessment of each candy's value.

Final Selection Using Rank Sum Method

To combine the results of these three analyses, the Rank Sum Method was applied. Each candy was assigned a rank in each of the three analyses, and those ranks were summed to determine the final top three candies. The candies with the lowest total rank sum were selected as the winners, as they consistently performed well across all analyses.

The final top three candies based on the Rank Sum Method were:

  1. Snickers – With a total of 11 points, Snickers performed consistently well across all three analyses, making it the top overall candy.
  2. Twix – With 15 points, Twix secured high rankings in all analyses, earning its spot as a top candy.
  3. Reese’s Peanut Butter Cup – Accumulating 16 points, Reese’s Peanut Butter Cup demonstrated strong consumer appeal and favorable attributes.

This method ensured a fair and balanced evaluation, incorporating both consumer preferences and a detailed analysis of key candy attributes to select the best Halloween treats.

Additional project images

Discussion and feedback(5 comments)
comment-1941-avatar
Kevin Webster
Kevin Webster
6 days ago
What if I want hard or fruity candy?

comment-1944-avatar
Andy Behar
Andy Behar
Project owner
5 days ago
Project owner
hello Kevin, hard and fruity candies didn’t score as well in this analysis, but feel free to enjoy your favorites. Appreciate you bringing a different flavor to the discussion.

comment-1965-avatar
Jasmin Simader
Jasmin Simader
2 days ago
I really love the detailled analysis you did! Great work!

comment-1972-avatar
Dirk Strauß
Dirk Strauß
2 days ago
Great work, my favorite is the final scatter plot that shows which metrics indicate the most popular treats!
2000 characters remaining
Cookie SettingsWe use cookies to enhance your experience, analyze site traffic and deliver personalized content. Read our Privacy Policy.