Data Analysis

The Data Analysis Stage

• The people working with datasets may not be at all connected to the data they’re working with, but that data may still represent human populations. Remember you can’t ask for the personal opinion of every individual in a database of a hundred million photos.
• Different data analysis methods may give completely different results, comparing competing models is important, since it may reveal unexpected things about your data

This is when we actually crunch the numbers. It is important to remember that who is doing the analyzing is relevant to what they might see in the results.

Remember that just because it can be done does not mean that it should. Even if potentially profitable.

Food For Thought

• In an age where (in many parts of the world), people can get online and connected to the internet and all its data for next to no cost, how do we decide what classifies good data-science?

​

• What happens if two different methods of analyzing the same data give totally different results? What process do you follow to figure out which model to trust?

​

• Who writes the best-practices “rulebook” for data science, when the field is completely independent of states and borders? Who enforces standards before the cases end up in the courts?

​