Classification Game

Your challenge is to create a decision tree that you think can successfully classify properties as either being in New York or San Francisco. Your decision must be based upon a limited number of characteristics.

Credit: Dataset and original concept courtesy of Stephanie Yee and Tony Chu @ r2d3.us where you can find further reading on machine learning that extends this exercise.

Building your decision tree

In order to help build your decision tree, you have been provided a training set with details of 10 properties in San Francisco and 10 in New York. Use the data to decide on your factors and boundaries that you think will solve the challenge. You can also download the training set as CSV if you wish to do some analysis of the data in Excel. You can also drag and drop the cards around the screen.

Buttons

Run model: Try your tree with the cards.

Reset: Moves the cards back into the panel after an evaluation.

Evaluating your decision tree

Use the evaluation set to see if your tree works. You can select your own set of evaluation cards or generate 20 random cards.

Good luck!