Types of Data Science Questions?

By: Emiley J Printer Friendly Format

descriptive analysis

Goal->Describe a set of data: Commonly applied to census data,
eg. http://books.google.com/ngrams

explaratory analysis

Goal->Find relationships you dont know - discovering new connections, correlation doesn not imply causation
eg. www.sdss.org/

inferential

Goal->Use a relatively small sample of data to say something about a bigger population - common statistical model
eg. effect of air polution control on life expectency in US from analysing 545 counties

predective

Goal->Use data on some objects to predict values for another object - x predicts y doesnot mean x causes y
eg. http://fivethirtyeight.blogs.nytimes.com
eg. how target figured out teen girl was pregnant

causal

Goal->To find out what happens to one variable when you make another variable change - usually randomized studies (gold standard for data analysis)
eg. Like a what-if analysis

mechanistic

Goal->Understand the exact changes in variables that lead to changes in other variables for individual objects - physical / engineering science
eg. Mechanistic - Empirical pavement design - fhwa.dot.gov

Data is the second most important thing in data science. the first most important is the question that you want to solve.