Our second Power BI task had us looking at a Crossword Dataset. The data was mostly clean, only requiring a custom key column for the capability of joining the puzzle information and answer information. I decided to focus solely on The New York Times and answers for this task, to find out what answers are the most common and how long those words typically are.
My main struggles with this task were finding valuable insights from the data to draw a story from. Once I had finished analysing the answer length and most common words, I began searching word APIs and reading interviews and articles on crossword editors and authors. I stumbled across the Datamuse API, which contained an endpoint that can count how many syllables are in words. I used this call to count how many syllables were in the answer column. Due to the size of the dataset, I narrowed it down to only 2019.
Answer Length
Main findings with answer length was that ‘S’ was the most common starting character, making up for 466 thousand of the answers. When it comes to word length, four letter words are the most common (859k), followed by five letter words (817k). Highlighting a clear favourability for short words, likely due to the constraints of a crossword grid.
Syllables and Clues
Since most answers are short, mostly ranging from four and five letter words, syllable counts are low, with one syllable words being the most common. Also, there is a pattern where longer answers require longer clues, suggesting that more complex words require more explanation.
Palindromes
There is a total of 13.3k Palindromes since 2001, only making up 1.76% of the total answers. They are most commonly short words, with three letter answers making up 78.2% of the total. These palindromes are mostly abbreviations, with the clues provided primarily being fill in the blank style.
