Questions
Discuss among the team and come up with five distinct question classes relevant to the Yelp dataset.
Here is a simple question class provided as an example that has only one variable.
How many businesses have more than X review counts?
Three example of specific questions in this class are:
- How many businesses have more than 1000 review counts?
- How many businesses have more than 500 review counts?
- How many businesses have more than 100 review counts?
Come up with five additional question classes. Each must have two or three variables.
Location
Three example of specific questions in this class are:
- Trendy vs. Location (What places have the most "trendy" places?)
- Hipster vs. Location (What's the most hipster place?)
- Just location by count (Where are the most popular places to do Yelp review?)
Stars
Three example of specific questions in this class are:
- Stars vs States (Are some states harsher on grading?)
- Stars vs Kid Friendly (Are family friendly restaurants given a more leannant rating?)
- Stars vs Hipster (Are hipsters harsher critics?)
Number of reviews vs number of stars
Three example of specific questions in this class are:
- (specific question)
- (specific question)
Noise Level
Three example of specific questions in this class are:
- Noise Level vs. Good for Kids (Correlation between kids and noise?)
- Noise Level vs. Romantic (Does noise level affect romance?)
- Noise Level vs. Type Of Restaurant (Are noise levels are only present in restaurant reviews?)
Good For Dancing
Three example of specific questions in this class are:
- Dancing vs Romance (How many couples prefer to dance on a date?)
- Dancing vs Rating (Will places get dinged if they allow dancing?)
- Dancing vs Trendy (Is dancing trendy? (If we had more data such as the type of dancing, could we predict what dancing is more trendy?))