Objective 2.2 Clean Data Sets CERTIFIED-DATA-SCIENCE-PRACTITIONER Practice Questions (Page 8)

Questions 36–40

1. 36foundation · easy

   What is the primary risk of leaving duplicate records in a dataset?

   • AIt increases the storage cost of the data warehouse
   • BIt violates data privacy regulations
   • CIt can bias statistical analysis and model training
   • DIt makes the data harder to read for humans
   Reveal Solution Discussion Select an answer first
2. 37foundation · easy

   Which type of bias occurs when the data labeling process is influenced by the labeler's own beliefs?

   • ASampling bias
   • BBandwagoning bias
   • CObserver bias
   • DConfirmation bias
   Reveal Solution Discussion Select an answer first
4. 38foundation · easy

   What is the role of data governance in data cleaning?

   • ATo increase the speed of cleaning
   • BTo define standards and policies that cleaning processes must follow
   • CTo eliminate the need for data cleaning
   • DTo automate all cleaning steps
   Reveal Solution Discussion Select an answer first
5. 39application · medium

   A data scientist is working with a dataset of user interactions. The data was collected from two different platforms, and a user can have multiple sessions. The 'session_id' is unique per platform, but a user can have the same 'session_id' on both platforms. The goal is to analyze user behavior across platforms. What is the most appropriate way to deduplicate this data?

   • AKeep all rows, as each interaction is a unique event.
   • BRemove all rows from one of the platforms to avoid confusion.
   • CRemove duplicate rows based on the 'session_id' column.
   • DCreate a composite key from 'user_id' and 'session_id' to identify unique sessions.
   Reveal Solution Discussion Select an answer first
6. 40expert · hard

   A model is being developed to screen job applicants. The team is concerned about bias against applicants from non-English-speaking backgrounds. The team has decided to use a fairness metric to evaluate the model. Which of the following approaches would be the most comprehensive way to evaluate and mitigate bias?

   • ARemove all features related to language or nationality from the model.
   • BRely on a single fairness metric, such as demographic parity.
   • CUse only the overall accuracy of the model to ensure it is high.
   • DEvaluate the model using multiple fairness metrics and compare them across groups.
   Reveal Solution Discussion Select an answer first