Marauder's Map: Exploring the Wizarding World of Data

Using one's creativity and data analysis skills, the goal of this challenge is to capture the magic of the beloved series on a single-page visual. The dataset contains 6 tables consisting of the Dialogue, Chapters, Movies, Characters, Places, and Spells tables.

Shaping the Data

Aside from the given columns and measures that participants can develop amongst them, I also wanted to explore who among the characters cast the most spells. Looking at the Spells and Dialogues tables, however, there is no way to merge these tables or use one to lookup columns in the other since there were no matching columns.

Using Power Query, I needed to essentially "extract" the different incantations within each dialogue. Here are the steps for the text analysis:

1. Split the Dialogues column into rows which contains sentences delimited by periods, exclamations points, and question marks.
2. Merge Dialogues table into the Spells table using Left Outer Join to return all from the first table (Dialogues) and their corresponding matches from the second table (Spells).
3. Use Fuzzy matching options to merge the tables. Play with the Similarity Threshold and see which option returns the most matches but not compromising its accuracy. I personally decided to go with a Similarity Threshold equal to 0.55.
4. Check the resulting table's matching accuracy and remove rows with incorrect matching, such as Nox with Not/Now/No, and Finite with Finish/Minute. Remove rows with null matches as well optimize the model.

In conclusion, this data challenge is an exciting opportunity for data analysts to showcase their creativity and data analysis skills while diving into the world of the beloved series. The final dashboard will not only provide insights into the world of Harry Potter but also serve as a testament to the participants' skills and knowledge. May the magic be with you on this journey!