Data Cleaning and Standardization PCAD-31 Practice Questions (Page 1)

Questions 1–5

1. 1application · medium

   A healthcare analyst is comparing patient recovery times across two hospitals. Hospital A records recovery time in days (ranging from 1 to 30), and Hospital B records it in weeks (ranging from 0.14 to 4.29). The analyst needs to combine these into a single dataset for a comparative study. What is the most critical preprocessing step?

   • AMerge the datasets as-is and rely on the model to handle the unit difference.
   • BConvert all recovery times to a single unit (e.g., days) before merging.
   • CApply z-score standardization to each hospital's data separately before merging.
   • DApply min-max scaling to each hospital's data separately before merging.
   Reveal Solution Discussion Select an answer first
2. 2expert · hard

   A data scientist is analyzing sensor data from a manufacturing plant. The 'temperature' column has a few extreme values that are physically possible but rare (e.g., during a machine malfunction). The 'vibration' column has a value of -1, which is physically impossible. The scientist needs to build a robust model that is not skewed by these values. The dataset is small, and every row is valuable. What is the most appropriate strategy?

   • AImpute both the impossible vibration value and the extreme temperature values with the median.
   • BApply min-max scaling to both columns, which will compress the outliers.
   • CTreat the impossible vibration value as erroneous and impute it, then apply a robust scaling method like RobustScaler to handle the extreme temperature values.
   • DRemove all rows with any extreme values in either column.
   Reveal Solution Discussion Select an answer first
4. 3application · medium

   A retail analytics team receives a CSV export from a legacy point-of-sale system. The 'transaction_date' column contains values like '2024-03-15', '03/15/2024', and '15-Mar-2024'. The 'store_id' column contains 'NYC-01', 'nyc-01', and 'NYC01' for the same physical store. The team needs to merge this data with a clean CRM dataset for time-series analysis. What is the most appropriate first step?

   • AApply min-max scaling to the date values to bring them into a 0-1 range.
   • BDrop all rows where the date format is not 'YYYY-MM-DD' to ensure the dataset is uniform.
   • CConvert the date column to a string and leave the store_id column as-is, since the merge will still work.
   • DConvert all date values to a single ISO 8601 format (YYYY-MM-DD) and map all store_id variants to a canonical value.
   Reveal Solution Discussion Select an answer first
5. 4foundation · easy

   What is the purpose of imputation in data cleaning?

   • ATo scale features to a common range
   • BTo replace missing values with estimated or substituted values
   • CTo remove duplicate rows from the dataset
   • DTo identify outliers in the data
   Reveal Solution Discussion Select an answer first
6. 5application · medium

   A logistics company has a dataset of shipment records. The 'weight_kg' column contains values like '12.5', '12,5', and '12.5kg'. The 'status' column contains 'Delivered', 'delivered', and 'DELIVERED'. The analyst needs to compute the average weight of delivered shipments. What is the most appropriate sequence of steps?

   • AFilter for 'Delivered' first, then compute the average of the 'weight_kg' column as-is.
   • BRemove all rows where 'weight_kg' is not in the format '12.5'.
   • CConvert the 'weight_kg' column to a string and concatenate it with the 'status' column.
   • DStandardize the 'weight_kg' column to a numeric type and the 'status' column to a consistent case, then filter and compute the average.
   Reveal Solution Discussion Select an answer first