AI in Data Science Workflows

End-of-lesson check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-ai-coding-AI-and-data-science-workflows-creators

1. What is a primary advantage of using AI to accelerate Exploratory Data Analysis (EDA)?

   1. AI can automatically suggest relevant visualizations based on data patterns
   2. AI can replace the data scientist entirely during EDA
   3. AI eliminates the need to examine data quality issues
   4. AI can complete EDA without any human oversight

2. Which task in a data science workflow is LEAST appropriate for AI automation?

   1. Suggesting appropriate visualization types
   2. Synthesizing findings into a summary report
   3. Generating initial model code drafts
   4. Determining which business question the analysis should address

3. What aspect of a data science project must remain under the data scientist's authority despite AI assistance?

   1. Formatting visualization axis labels
   2. Writing documentation comments
   3. Selecting which libraries to import
   4. Choosing substantive decisions based on domain expertise

4. When AI generates model code automatically, what is the data scientist's primary responsibility?

   1. Deleting any code AI produces to maintain control
   2. Reviewing and validating the code for correctness and appropriateness
   3. Submitting AI-generated code directly to production
   4. Typing the code manually since AI cannot be trusted

5. Why is domain knowledge considered irreplaceable in data science projects?

   1. Because AI systems cannot read documentation
   2. Because regulations prohibit AI from making decisions
   3. Because substantive choices require understanding context that AI lacks
   4. Because data scientists are cheaper than AI tools

6. A data scientist uses AI to generate multiple model candidates, then selects the most appropriate one based on business constraints. This demonstrates what principle?

   1. AI should make all final decisions
   2. Technical choices should be fully automated
   3. AI has replaced the data scientist role
   4. Domain judgment remains central to substantive choices

7. What type of content can AI help synthesize for stakeholder reports?

   1. Statistical code comments
   2. Database schema designs
   3. Raw data dumps
   4. Executive summaries of findings

8. In stakeholder reporting, what limitation requires human intervention when AI assists?

   1. AI may not accurately interpret stakeholder priorities or organizational context
   2. AI cannot type words
   3. AI lacks access to the internet
   4. AI cannot format PDFs

9. What is required to determine whether an AI-augmented data science project achieved its goals?

   1. Checking how long the AI tool ran
   2. Counting how many lines of code AI generated
   3. Evaluating only the technical accuracy of AI suggestions
   4. Measuring outcomes against predefined success criteria

10. Why can AI not fully replace stakeholder relationships in data science?

    1. Because AI tools are too expensive
    2. Because stakeholders prefer slow responses
    3. Because trust, communication, and context understanding require human connection
    4. Because AI cannot attend meetings

11. Which AI capability directly accelerates the Exploratory Data Analysis phase?

    1. AI-suggested visualizations based on data characteristics
    2. AI writing the final research paper
    3. AI filing regulatory documents
    4. AI scheduling meetings with stakeholders

12. When AI suggests a visualization during EDA, what should the data scientist evaluate?

    1. How many colors the AI used
    2. What programming language the suggestion uses
    3. Whether the visualization accurately represents the data and answers the relevant question
    4. Whether the AI tool has a nice user interface

13. What is a key benefit when AI generates tests for model code?

    1. AI can quickly produce test cases covering common scenarios
    2. AI replaces the data scientist entirely
    3. AI guarantees the model will be perfect
    4. AI eliminates the need for any human testing

14. Even when AI generates model code, why does human review remain necessary?

    1. AI code is always secure and perfect
    2. Human review is required by law
    3. AI cannot generate code without errors
    4. Generated code may contain logical errors, inappropriate assumptions, or miss edge cases

15. What risk emerges from over-relying on AI during the modeling phase?

    1. Stakeholders will become too satisfied
    2. AI will demand higher pay
    3. The data scientist may lose engagement with critical modeling decisions
    4. The project will be completed too quickly