This might be a dumb question, but I am an absolute beginner.

What kind of data analysis problem is the following?

Is it an Exploratory Data Analysis problem/Data Mining problem/data analytics problem?

Secondly, what steps can I follow to solve the following problem? Is there any algorithm?

Records in this data set describe decks of cards used in a popular collectible card video game called "Clash Royale". These decks were obtained using the RoyaleAPI.com service, from games that took place in January 2019. Each record consists of five values:

  • timestamp of the game (column timestamp),
  • arena ID (column arena_id – higher the arena, more skilled/experienced a player is)
  • outcome of a game (column has_won, 1 – the player won, 0 the player lost)
  • a player ID (column tag)
  • list of exactly eight cards in the player’s deck separated by “_” signs (column player_deck)

Your task is to analyze this data and search for interesting card usage patterns, and interactions/dependencies between cards.

For example:

  • find card combos that were particularly popular in January 2019 (e.g., top 100 card sets with regard to their support, top 100 card sets of size 2, size 3, etc.),
  • identify those card combos which have high win-rates (e.g., top 100 card sets with regard to win-rate and with support > 1%),
  • does the card usage/popularity/effectiveness changes in time?
  • does the arena level have any influence on card usage/popularity/effectiveness?
  • find interesting associations between cards,
  • can you cluster players according to their play patterns and card preferences? Additionally, design and construct a card recommender system that allows players to indicate four cards which they want to have in a deck, and recommends the remaining four to create a reasonable deck. How can you evaluate the effectiveness of the recommendations? You have to report the discoveries in the form of an R notebook (with code and all computation outcomes). You have to remember about visualizations. I.e. you have to make this report as visually interesting for a reader as you can.

1 Answer 1


You should start with exploratory data analysis to understand the data in as much detail as possible. Once you have a good understanding of the obvious patterns in the dataset, you should then turn your focus towards answering the less obvious questions using modelling techniques.

The final item on the list that hints at building a recommender is an example of data modelling whereas all other items appear to be exploratory data analysis.

It appears this is an open ended problem. It is for you to look at the data, gain an understanding out of it and then figure out answers to interesting and less obvious questions that might interest the Business group.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.