Data cluedo: solving a murder mystery with data science


If you’re a lover of board games, then you probably know Cluedo or Clue, as it’s known in the U.S. Cluedo is a murder mystery game where players must use deductive reasoning to answer who the murderer is, which weapon they used and where the murder took place. ORTEC and EY teamed up to create a fun spin-off of the game that challenges players to solve a murder mystery using data science techniques. The game was played by Dutch students during the Amsterdam Data Challenge, a sideline competition of The Econometrics Game. The Econometrics Game is a prestigious global econometrics competition that brings students from top universities together to compete for a prize. The theme of this year’s Econometric Game revolved around crime and safety, so it was the perfect opportunity to put it to the test.

How does it work – from data scientists to detectives

Much like the original Cluedo, Data cluedo is all about putting different pieces of a puzzle together to solve a crime. The difference is that you get a lot more data to work with and you need some data science expertise to play. Participants get several datasets in CSV files which can be read in R. Once they load the data into the software package, they need to use different statistical techniques to answer 5 Ws:

  • Why did the murder take place.
  • Where did it take place.
  • When did it happen.
  • Who committed the crime.
  • What did the murderer use as a weapon.

Six teams of students were the first to play Data cluedo at the Amsterdam Data Challenge. The teams were given 5 data sets, ranging from a long list of emergency phone calls made to the police, to a large chunk of real Tweets which were slightly modified to add even more excitement to the game.

Teams had to perform sentiment analysis on the given Twitter posts to quantify and analyze different emotions in order to understand the intent of potential murderers. They also had to employ a machine learning algorithm on different chemical compounds found at the scene to predict which weapon was used.

The Winning Team

From left to right: Xander van den Berg (EY),  Casper Hoogenboom, Wietse Steenstra, Eveline Duyster And Sanne van Gellekom.

After a thrilling day full of riddles, The Unusual Suspects, a team from the University of Amsterdam, was declared the winner of Data cluedo. Participating teams were judged based on how good their storyline was, how close they came to the original storyline and how good their presentation was. The Unusual Suspects was made up by Sanne van Gellekom, Eveline Duyster, Casper Hoogenboom and Wietse Steenstra. As winners, they got free tickets to watch a football game in luxury at the ORTEC Sky Box in the Amsterdam Arena – home to Ajax, the legendary Dutch football team.

A fun concept for companies

Data cluedo is great for students who want to put their skills into practice, but it can also be very useful for companies. Businesses and organizations can use it for recruitment and employee training, as it offers a fun way to learn new data science techniques, practice communication skills, and stay on top of new developments.

Data cluedo was created by Mor Verbin (OR Engineer and Data Scientist at ORTEC), Ivo Fugers (Data Scientist at ORTEC) Xander van den Berg (Senior Advisor Data Analytics at EY), Mariska van Ham (Advisor Actuarial Services Data Scientist at EY), and Twan van de Waerdt (Senior Advisor Data Analytics at EY).

Contact us to learn more about the game.

· · · · ·

Leave a Comment