Leveraging Python GMLE Practice Questions (Page 4)

Questions 16–20

1. 16application · medium

   A data scientist at a retail company loads a CSV of daily sales into a Pandas DataFrame. The 'revenue' column contains some missing values and a few non-numeric strings like 'N/A'. The team needs to compute the mean revenue per store, ignoring missing values, and then train a linear regression model using the cleaned data. Which sequence of operations should be used?

   • Adf = df[df['revenue'].str.isdigit()]; df.groupby('store')['revenue'].mean()
   • Bdf['revenue'] = pd.to_numeric(df['revenue'], errors='coerce'); df = df.dropna(subset=['revenue']); df.groupby('store')['revenue'].mean()
   • Cdf['revenue'] = df['revenue'].astype(float); df.groupby('store')['revenue'].mean()
   • Ddf['revenue'] = df['revenue'].fillna(0); df.groupby('store')['revenue'].mean()
   Reveal Solution Discussion Select an answer first
2. 17foundation · easy

   Which Pandas method is used to check for missing values in a DataFrame?

   • Adf.fillna()
   • Bdf.isnull()
   • Cdf.clean()
   • Ddf.dropna()
   Reveal Solution Discussion Select an answer first
4. 18expert · hard

   A data engineer has a DataFrame with a column 'timestamp' containing datetime strings in the format 'YYYY-MM-DD HH:MM:SS'. They need to filter rows where the timestamp is in the year 2023. They are considering using string operations versus converting to datetime. Which approach is more robust and efficient?

   • AUse df[df['timestamp'].apply(lambda x: x.startswith('2023'))]
   • BUse df[df['timestamp'].str.contains('2023')]
   • CUse string slicing: df[df['timestamp'].str[:4] == '2023']
   • DConvert the column to datetime using pd.to_datetime and then filter with df[df['timestamp'].dt.year == 2023]
   Reveal Solution Discussion Select an answer first
5. 19application · medium

   A machine learning engineer is building a binary classifier to predict customer churn. The dataset has 10,000 rows and 50 features, including categorical variables like 'contract_type'. They want to use a logistic regression model. Which pipeline correctly preprocesses the data and trains the model?

   • ALogisticRegression directly on the raw DataFrame
   • BLabelEncoder for all features, then LogisticRegression
   • CStandardScaler on the entire DataFrame, then LogisticRegression
   • DOneHotEncoder for categorical features, StandardScaler for numeric features, then LogisticRegression in a ColumnTransformer
   Reveal Solution Discussion Select an answer first
6. 20application · medium

   A data scientist has a DataFrame with a column 'price' that contains outliers. They want to replace outliers beyond 3 standard deviations from the mean with the median. Which code correctly performs this winsorization?

   • Amean = df['price'].mean(); std = df['price'].std(); median = df['price'].median(); df.loc[(df['price'] - mean).abs() > 3*std, 'price'] = median
   • Bdf['price'] = df['price'].clip(lower=mean - 3*std, upper=mean + 3*std)
   • Cdf['price'] = df['price'].replace(df['price'][(df['price'] - mean).abs() > 3*std], median)
   • Ddf['price'] = df['price'].apply(lambda x: median if abs(x - mean) > 3*std else x)
   Reveal Solution Discussion Select an answer first