The Data Science Workflow · seed 1 · A4, ink-friendly. The answer key prints on its own page for grown-ups.

From vague question to honest answer

Computing · Data & Databases · ages 21-22
Name ______________________   Date ____________

  1. Which rewrite can order records actually answer?

    • Which pastry smells nicest
    • Which new customers stopped after one purchase
    • Why the sky is blue

  2. Which of these questions could you answer by collecting data?

    • Which snack is most popular in our class?
    • What is the best snack in the whole world?
    • Should snacks be allowed in class?

  3. Your boss asks: how do our shop's sales compare with sales across the whole country? Your notebook holds only your own shop's sales. Which part of that question can your data answer?

    • A summary of this shop's sales only
    • The total sales of every shop in the country
    • The average sales of all shops in the country
    • Next year's sales for every shop

  4. Which order matches the analysis loop?

    • Question, data, build, look, report
    • Report, build, data, question, look
    • Build, report, look, question, data

  5. Your course grade counts homework for 20 percent, quizzes for 30 percent, and the final test for 50 percent. You scored 80 on homework, 90 on quizzes, and 70 on the test. What is your final grade?

    Answer: ______________

  6. What keeps an analysis reproducible?

    • Steps on file for anyone to rerun
    • Deleting the raw data after plotting
    • Guessing the numbers from memory

  7. Five friends scored 4, 5, 5, 6, and 20 on a quiz. What is the median score?

    Answer: ______________

  8. A vague churn ask becomes answerable when it names what?

    • A bigger font for the report
    • Who, what, and which records
    • A new brand color

  9. A till upgrade logs items starting next month. What can you honestly do today?

    • Rank pastries from daily totals anyway
    • Invent item numbers that look plausible
    • Report sums now, study items later

  10. Two analysts get different numbers from the same data. What do you check first?

    • Whose laptop is newer
    • Whether steps differ
    • Who joined the team first
LightMySky · lightmysky.comW1-mt_99QNfiyQ55-s1

Answer key

For grown-ups. Fold this page away before handing over the rest.

From vague question to honest answer W1-mt_99QNfiyQ55-s1

  1. Which new customers stopped after one purchase · Order records hold dates and repeat counts, which show exactly that behavior.
  2. Which snack is most popular in our class? · A question data can answer is about something you can count or measure for many cases. Opinions about the whole world or rules you want changed are not countable that way.
  3. A summary of this shop's sales only · Summarizing the data you actually have is a descriptive question. Reaching beyond it to the whole country would need data you do not have.
  4. Question, data, build, look, report · You state the question first and report so someone can act last.
  5. 78 · Multiply each score by its weight and add: 16 + 27 + 35 = 78. Each piece of data only counts as much as its share allows.
  6. Steps on file for anyone to rerun · Written steps let a second person rerun everything and land on the same numbers.
  7. 5 · The median is the middle value after ordering, which is 5. The mean would be 8, pulled up by the single 20, so the median is the summary this data can honestly support.
  8. Who, what, and which records · Group plus behavior plus data turns a worry into a question the records can answer.
  9. Report sums now, study items later · Today's data bounds today's claims, while next month's data opens the item question.
  10. Whether steps differ · Same data plus different steps explains different numbers, so compare the records.
Worksheet · LightMySky