# Possible Challenges for a Data Science Escape Room

Dear Data Science Community,

as my project for my bachelorthesis I am working on a concept for a physical Data Science Escape Room. The goal of the escape room is to create awareness for data science / the importance of data in general and motivate "escapees" to learn more about data science.

The escape room is going to be a physical room, that needs to be escaped by completing different challenges related to data science. The target group are people who are already working with data in form of reports or similar, but have no or little to none background in data science.

Possible challenges for this escape room could use simple machine learning algorithms that need to be trained, or easy statistical concepts. Players could also need to curate data, which then needs to be integrated to be usable afterwards.

I am looking for ideas or further input for possible challenges for this escape room, and was hoping to have some data science experts give creative input!

Thank you!

Fun! Cool idea, but it seems like a big undertaking. Anyways, here is my suggestion:

Summary

A manual fit of a 2D logistic regression which only becomes solvable once some previous step is completed.

Explaination

I imagine being greeted by a screen looking something like this:

The two boxes are inputs where the escapees can input numbers. If they try inputting at random it will be possible to figure out that the numbers represent $$B_0$$ and $$B_1$$ in: $$Y = B_0 + B_1 X1$$

This will still leave them confused since they will only see something like this:

However, when some other challenge in the room is finished the real challenge is revealed by colorizing the dots.

Then they just need to find the proper intercept and coefficient to finish challenge.

I think this would work quite well if there are enough samples so that you won't be able to hack it by trying values before they know the colors. If the challenge is too easy you could make them fit polynomials instead.

• Thanks alot, I really like your idea. I love how tangible it is, and that its kind of a two step progress! – Yannick Ruettgers Apr 11 '19 at 7:33
• Glad you liked it! If you ever implement it I would love it if you would drop a comment here and let me know. :) – Simon Larsson Apr 11 '19 at 7:41

### Suggestion 1: Humans vs Machines

You could try showing the power of Data Science by having they compete with machines in tasks related to their field or with day-to-day life (for example classifying images from image-net).

### Suggestion 2: A Human Brain of Humans

You could also separate then into groups (layers) and suppose every person is a neuron into a neural network. They are supposed to guess the input label without ever seen the entire input. (I am using Mind Field Episode The Stilwell Brain as inspiration)

• I like your ideas as well. Suggestion 1: Im afraid that it could scare them more of machines taking their jobs than making them interested in the topic. But I think I can work something out there. Sugg. 2: Thats a good Idea, is it worth to try youtube premium to watch the episode? – Yannick Ruettgers Apr 11 '19 at 7:41
• I watched the whole series with my YouTube Premium Free trial and I say to you: Mind Field is worth every second hahah – Pedro Henrique Monforte Apr 11 '19 at 10:11
• You should make tem compete with ANI and show them there is no AGI. That actually AI is helpful but dependent on us. Then, after que competition you can make then colaborate and show how tasks that we usually find boring can be lifted by AI and enable us to do the fun, human and creative part of our jobs . – Pedro Henrique Monforte Apr 11 '19 at 10:15
• Okay, ill give it a go! :) Thats a good idea, thanks! – Yannick Ruettgers Apr 11 '19 at 11:46