- | Page
CIS 5450 - MIDTERM 1|| QUESTIONS AND
100% ACCURATE ANSWER S.
Question : How is data science different from fields like physics
and the rest of engineering?
Correct answer: Other fields tend to be model first, while data
science tends to be simulation first
Question : What are the four steps to the data science process?
Correct answer: Raw Data => Structured, Cleaned Data =>
Extracted Features => Descriptive or Inferential Model
Question : What makes big data "big"?
Correct answer: - Too complex for a human to understand
Doesn't fit into a single memory space
Requires multiple computers
High dimensionality
High velocity
- | Page
Question : What is the difference between raw data and
structured data?
Correct answer: Structured data is in a form that analysis can be
done - machine learning features or fields, objects etc.
Question : What is the difference between the data science myth
and reality?
Correct answer: The data science myth is that we can just turn
the crank and use fancy machine learning algorithms to go from nothing to insights. The reality is that we rely on human experts to choose the right models and features.
Question : What are questions we ask in the data science
application process?
Correct answer: - What question are you answering?
What is the right scope of the project?
What data will you use?
- | Page
What techniques are you going to try?
How will you evaluate your results?
What maintenance will be required?
Question : What are the three basic encodings of structured
data?
Correct answer: Table (of tuples), Arrays, Nested document
(json)
Question : What function can be used to extract a table from
html?
Correct answer: pandas.read_html (pointed to a url)
Question : How is information extraction defined?
Correct answer: Taking information from unfiltered and
unprocessed text sources (such as webpages)
Question : What is the relational algebra operation known as
projection?
- | Page
Correct answer: Essentially filtering to one column. As opposed
to selection, which filters to a certain subset of rows. Ex: We
project the CEO name column, clean it, and store it "clean_names"
Question : How would you perform a function on every item in
a row or column in a dataframe?
Correct answer: df.apply(f)
Question : Under what circumstances would a join/merge put
together two rows?
Correct answer: If the selected attribute was the same in both
rows.
Question : What keyword does SQL use to combine two
databases?
Correct answer: JOIN
Question : What is the difference between a left and right join?
Correct answer: A left join will keep all items in the left table
even if they don't have a match in the right table. The opposite is