Chapter 5 Magic data

Magic: The Gathering is a popular collectible card game that includes strategy and chance. If you have not played, here is a great description. This is an excellent example of data collected that is appropriate for experimentation for several reasons. Many games have expansion packs and add-ons. Physical card collecting from sports to fantasy is a very popular activity in addition to a multi-million dollar economy and sales continue to grow. Consequently, these data are an opportunity to explore likelihood and consumer buying choices. It is also fun. Incidentally, the most valuable card reportedly sold for $250,000USD so understanding how frequently rare cards pops up is pretty interesting.

Learning outcomes

  1. Work with an existing dataset and reuse it.
  2. Design questions from data.
  3. Connect principles of experimental design to implementation with data.
  4. Write a clear hypothesis and predictions to explain or predict patterns in data.
  5. Communicate data, design, and science succinctly.

Steps

  1. Download the dataset.
  2. Read the meta-data, and review what the game involves.
  3. Explore likelihood topics, and prep a list of questions for the data.
  4. Test one question with the data via a plot and a statistical test.
  5. Decide if this is the dataset for you to write up as short research note.

Data

Here are the data collected by a researcher at York University. These data are the first of their kind and super fascinating.

Deeper dive

If you choose this adventure, your goal is to explore any component of the data to apply design principles. Informally, we are data mining (Hand et al. 2000). Data mining is a field of data science and many other disciplines that is exploding in capacity and application (Hooshyar, Yousefi, and Lim 2019; Aldowah, Al-Samarraie, and Fauzy 2019). Typically, mining data includes seeking patterns and building models that either focus on description or on prediction. Experimental design principles are not always a component of data mining endeavours, and this is unfortunate. We can do better science and build better models with design thinking principles. It is still an experiment, but someone else (or you) collected data. Data can be used for the explicit purpose they were collected, re-purposed to explore another idea, or simply collected without a priori hypotheses and predictions that we derive later. This is where design thinking can make profound contributions. Select a variable from the data that you think can be a meaningful mechanism or explains patterns, describe differences, or predict outcomes. In this example, this design and experimentation process can include good thinking and data use to examine price prediction, likelihood of rares or other card categories, and pattern frequencies by various card attributes. Begin with one key variable that can be a factor or grouping variable and one key variable that is a response or outcome. Then, do some work, some thinking, try some designs with the data, and make the call if this is data-design lab you will write up.