Python for Data Science Practice Exam

Question 1 Python for Data Science

How do you create a 3x3 identity matrix in NumPy?

df['date_col'].dt.year uses the .dt accessor to extract the year component. np.vstack([a, b]) stacks arrays on top of each other along axis 0. It reduces memory usage and speeds up operations like groupby on columns with a limited number of unique values. np.eye(3) returns a 3×3 array with ones on the diagonal and zeros elsewhere.

Question 2 Python for Data Science

What does `pd.merge(df1, df2, on='key')` do by default?

df.reset_index(drop=True) resets the index to default integers; drop=True discards the old index. It performs an inner join, keeping only rows where the key exists in both DataFrames. Press A in command mode (press Esc first) to insert a cell above. df.groupby('group')['val'].transform(lambda x: (x - x.mean()) / x.std()) standardizes values within each group.

Question 3 Python for Data Science

Give an example of NumPy broadcasting with a scalar.

df.reset_index(drop=True) resets the index to default integers; drop=True discards the old index. Press A in command mode (press Esc first) to insert a cell above. df.dropna(how='all') removes rows where every column value is NaN. arr * 3 multiplies every element of arr by 3 without an explicit loop — the scalar is broadcast across the array.

Question 4 Python for Data Science

How do you apply a custom function element-wise to a column?

np.random.randint(0, 100, size=(3, 4)) creates a 3×4 array of random integers between 0 and 99. It performs an inner join, keeping only rows where the key exists in both DataFrames. df['col'].apply(lambda x: x ** 2) applies the lambda function to each element of the column. Shift + Enter executes the current cell and advances to the next cell.

Question 5 Python for Data Science

How do you sort a DataFrame by multiple columns?

Use plt.title('Title'), plt.xlabel('X'), and plt.ylabel('Y') before calling plt.show(). df.sort_values(['col1', 'col2'], ascending=[True, False]) sorts by col1 ascending, then col2 descending. arr[1, 2] uses zero-based indexing to retrieve the element at the second row, third column. np.random.randint(0, 100, size=(3, 4)) creates a 3×4 array of random integers between 0 and 99.

Question 6 Python for Data Science

How do you save a matplotlib figure to a file?

plt.savefig('plot.png', dpi=300, bbox_inches='tight') saves the current figure as a PNG file. plt.scatter(x, y) plots individual data points as dots on the chart. df['date_col'].dt.year uses the .dt accessor to extract the year component. int64 for integers and float64 for floating-point numbers; use df.dtypes to inspect.

Question 7 Python for Data Science

What does `df.info()` display?

df[df['age'] > 30] returns a filtered DataFrame using boolean indexing. df.set_index('col') makes the specified column the new row index. df['total'] = df['price'] * df['quantity'] creates a new column using vectorized arithmetic. It prints the DataFrame's column names, non-null counts, data types, and memory usage — useful for a quick data overview.

Question 8 Python for Data Science

How do you convert a column to lowercase strings in Pandas?

df.loc[0:3, ['name', 'age']] selects rows 0–3 and only the 'name' and 'age' columns. np.linspace(0, 1, 5) returns array([0., 0.25, 0.5, 0.75, 1.]). Shift + Enter executes the current cell and advances to the next cell. df['col'].str.lower() uses the .str accessor to apply the lower() string method.

Question 9 Python for Data Science

How do you compute the dot product of two NumPy arrays?

np.dot(a, b) or a @ b computes the matrix/dot product of arrays a and b. np.random.randint(0, 100, size=(3, 4)) creates a 3×4 array of random integers between 0 and 99. plt.plot(x, y) followed by plt.show() draws a basic line plot. df.drop_duplicates() removes duplicate rows, keeping the first occurrence by default.

Question 10 Python for Data Science

How do you select specific rows and columns by label in Pandas?

It returns array([0, 2, 4, 6, 8]) — values from 0 up to (not including) 10 with step 2. df.loc[0:3, ['name', 'age']] selects rows 0–3 and only the 'name' and 'age' columns. It returns summary statistics (count, mean, std, min, 25%, 50%, 75%, max) for all numeric columns. df.duplicated() returns a boolean Series that is True for duplicate rows.

Question 11 Python for Data Science

How do you compute the correlation matrix of a DataFrame?

fig, axes = plt.subplots(1, 2, figsize=(10, 4)) creates a figure with 1 row and 2 columns of axes. df.dropna(how='all') removes rows where every column value is NaN. arr.reshape(3, 4) returns a 3×4 view of the array (total elements must match). df.corr() returns a DataFrame of pairwise Pearson correlation coefficients for all numeric columns.

Question 12 Python for Data Science

How do you extract the year from a datetime column?

df['col'].fillna(method='ffill') performs forward fill, propagating the last valid value. df['date_col'].dt.year uses the .dt accessor to extract the year component. np.eye(3) returns a 3×3 array with ones on the diagonal and zeros elsewhere. df.corr() returns a DataFrame of pairwise Pearson correlation coefficients for all numeric columns.

Question 13 Python for Data Science

What does `df.shape` return?

df.isnull().sum() returns the count of NaN values per column. A tuple (rows, columns) indicating the dimensions of the DataFrame. arr.reshape(3, 4) returns a 3×4 view of the array (total elements must match). df['col'] = df['col'].astype(int) casts the column to integer type.

Question 14 Python for Data Science

How do you reshape a 1D array of 12 elements into a 3x4 matrix?

arr.reshape(3, 4) returns a 3×4 view of the array (total elements must match). np.vstack([a, b]) stacks arrays on top of each other along axis 0. plt.savefig('plot.png', dpi=300, bbox_inches='tight') saves the current figure as a PNG file. pd.cut(df['age'], bins=[0, 18, 35, 60, 100], labels=['child', 'young', 'middle', 'senior']) creates categorical bins.

Question 15 Python for Data Science

How do you extract the day of the week from a datetime column?

df['date_col'].dt.dayofweek returns integers 0 (Monday) through 6 (Sunday). pd.read_csv('file.csv') reads the file and returns a DataFrame; use sep='\t' for TSV files. df[df['age'] > 30] returns a filtered DataFrame using boolean indexing. pd.merge(df1, df2, on='key', how='right') keeps all rows from df2 and matches from df1.

Question 16 Python for Data Science

How do you remove duplicate rows from a DataFrame?

Broadcasting is NumPy's ability to perform element-wise operations on arrays of different shapes by automatically expanding the smaller array to match dimensions. df.drop_duplicates() removes duplicate rows, keeping the first occurrence by default. df[df['age'] > 30] returns a filtered DataFrame using boolean indexing. It measures the wall time and CPU time for the entire cell's execution (single run, unlike %timeit).

Question 17 Python for Data Science

How do you perform element-wise comparison in NumPy?

df.loc[0:3, ['name', 'age']] selects rows 0–3 and only the 'name' and 'age' columns. df.set_index('col') makes the specified column the new row index. arr > 5 returns a boolean array where each element indicates whether the condition is True or False. df[df['age'] > 30] returns a filtered DataFrame using boolean indexing.

Question 18 Python for Data Science

How do you set a column as the DataFrame index?

df.set_index('col') makes the specified column the new row index. df.dropna(how='all') removes rows where every column value is NaN. df.loc[0:3, ['name', 'age']] selects rows 0–3 and only the 'name' and 'age' columns. df['col'].str.lower() uses the .str accessor to apply the lower() string method.

Question 19 Python for Data Science

How do you select the first 5 rows and first 2 columns by position?

df.iloc[:5, :2] uses integer-based indexing to slice rows and columns. fig, axes = plt.subplots(1, 2, figsize=(10, 4)) creates a figure with 1 row and 2 columns of axes. sns.pairplot(df) creates a grid of scatter plots for every pair of numeric columns, with histograms on the diagonal. pd.merge(df1, df2, on='key', how='outer') keeps all rows from both DataFrames, filling NaN where keys don't match.

Question 20 Python for Data Science

How do you add a new cell above the current cell in Jupyter?

pd.read_csv('file.csv') reads the file and returns a DataFrame; use sep='\t' for TSV files. df.groupby('group')['val'].transform(lambda x: (x - x.mean()) / x.std()) standardizes values within each group. Press A in command mode (press Esc first) to insert a cell above. pd.merge(df1, df2, on='key', how='left') keeps all rows from df1 and matches from df2.

Question 21 Python for Data Science

How do you check for missing values in a DataFrame?

Shift + Enter executes the current cell and advances to the next cell. It prints the DataFrame's column names, non-null counts, data types, and memory usage — useful for a quick data overview. It returns an array where positive values are kept and negative/zero values are replaced with 0. df.isnull().sum() returns the count of NaN values per column.

Question 22 Python for Data Science

How do you apply min-max scaling using scikit-learn?

pd.concat([df1, df2], axis=1) joins DataFrames horizontally, aligning on the index. sns.pairplot(df) creates a grid of scatter plots for every pair of numeric columns, with histograms on the diagonal. from sklearn.preprocessing import MinMaxScaler
scaler = MinMaxScaler()
df[['col']] = scaler.fit_transform(df[['col']]) It stacks DataFrames vertically (along axis 0) by default, appending rows.

Question 23 Python for Data Science

How do you access the element at row 1, column 2 of a 2D NumPy array `arr`?

df.sort_values('price', ascending=False) returns the DataFrame sorted from highest to lowest price. arr[1, 2] uses zero-based indexing to retrieve the element at the second row, third column. df.head(10) returns the first 10 rows (default is 5 if no argument is given). df.rename(columns={'old_name': 'new_name'}) returns a DataFrame with the specified columns renamed.

Question 24 Python for Data Science

What is broadcasting in NumPy?

Broadcasting is NumPy's ability to perform element-wise operations on arrays of different shapes by automatically expanding the smaller array to match dimensions. df.groupby('category')['value'].mean() computes the mean of 'value' for each unique category. df['col'] = df['col'].astype('category') converts the column to the memory-efficient category dtype. df['col'].nunique() returns the count of distinct values, excluding NaN by default.

Question 25 Python for Data Science

What does `df.dropna()` do?

np.eye(3) returns a 3×3 array with ones on the diagonal and zeros elsewhere. df.groupby('col').agg(['mean', 'sum', 'count']) returns a DataFrame with multiple summary statistics. It returns True if value is NaN, None, or NaT (Not a Time). It removes rows (by default) that contain any NaN values from the DataFrame.

Question 26 Python for Data Science

How do you fill missing values with the previous row's value?

Use plt.title('Title'), plt.xlabel('X'), and plt.ylabel('Y') before calling plt.show(). loc selects by label (index/column names), while iloc selects by integer position. pd.merge(df1, df2, on='key', how='outer') keeps all rows from both DataFrames, filling NaN where keys don't match. df['col'].fillna(method='ffill') performs forward fill, propagating the last valid value.

Question 27 Python for Data Science

What does the seaborn `pairplot` function do?

It returns array([0, 2, 4, 6, 8]) — values from 0 up to (not including) 10 with step 2. np.sqrt(arr) applies the square root to each element and returns a new array. sns.pairplot(df) creates a grid of scatter plots for every pair of numeric columns, with histograms on the diagonal. pd.concat([df1, df2], axis=1) joins DataFrames horizontally, aligning on the index.

Question 28 Python for Data Science

How do you interpolate missing values linearly in Pandas?

df['col'].interpolate(method='linear') fills NaN values using linear interpolation between existing data points. Broadcasting is NumPy's ability to perform element-wise operations on arrays of different shapes by automatically expanding the smaller array to match dimensions. df.head(10) returns the first 10 rows (default is 5 if no argument is given). It flattens a multi-dimensional array into a 1D array, returning a view when possible.

Question 29 Python for Data Science

What is feature scaling and why is it important?

Feature scaling normalizes the range of features so that algorithms sensitive to magnitude (e.g., KNN, SVM) perform correctly. Common methods: min-max and standardization. df.groupby('group')['val'].transform(lambda x: (x - x.mean()) / x.std()) standardizes values within each group. It removes rows (by default) that contain any NaN values from the DataFrame. pd.merge(df1, df2, on='key', how='left') keeps all rows from df1 and matches from df2.

Question 30 Python for Data Science

What is the difference between `transform` and `agg` in groupby?

It runs a statement multiple times and reports the average execution time, useful for benchmarking code. pd.merge(df1, df2, on='key', how='left') keeps all rows from df1 and matches from df2. agg returns a reduced result (one row per group), while transform returns a result with the same shape as the input. df['date_col'].dt.year uses the .dt accessor to extract the year component.

Question 31 Python for Data Science

What does `arr.T` do for a NumPy array?

df.corr() returns a DataFrame of pairwise Pearson correlation coefficients for all numeric columns. It returns True if value is NaN, None, or NaT (Not a Time). It returns the transpose of the array, swapping rows and columns. plt.savefig('plot.png', dpi=300, bbox_inches='tight') saves the current figure as a PNG file.

Question 32 Python for Data Science

How do you read a CSV file into a Pandas DataFrame?

np.eye(3) returns a 3×3 array with ones on the diagonal and zeros elsewhere. df[(df['age'] > 25) & (df['city'] == 'NYC')] — wrap each condition in parentheses and use & (and) or | (or). df['date_col'].dt.year uses the .dt accessor to extract the year component. pd.read_csv('file.csv') reads the file and returns a DataFrame; use sep='\t' for TSV files.

Question 33 Python for Data Science

How do you compute the element-wise square root of a NumPy array?

It creates a pivot table summarizing total sales by region (rows) and quarter (columns). df['col'].replace({'old_val': 'new_val'}) substitutes matching values in the column. It performs an inner join, keeping only rows where the key exists in both DataFrames. np.sqrt(arr) applies the square root to each element and returns a new array.

Question 34 Python for Data Science

What does `%matplotlib inline` do in a Jupyter notebook?

It configures matplotlib to display plots inline directly below the code cell. from sklearn.preprocessing import StandardScaler
scaler = StandardScaler()
df[['col']] = scaler.fit_transform(df[['col']]) fig, axes = plt.subplots(1, 2, figsize=(10, 4)) creates a figure with 1 row and 2 columns of axes. A ufunc is a function that operates element-wise on ndarrays, e.g., np.sqrt(arr), np.exp(arr).

Question 35 Python for Data Science

What is a universal function (ufunc) in NumPy?

df.reset_index(drop=True) resets the index to default integers; drop=True discards the old index. pd.DataFrame({'col1': [1, 2], 'col2': [3, 4]}) creates a DataFrame with two columns. pd.merge(df1, df2, left_on='id', right_on='user_id') joins using differently named key columns. A ufunc is a function that operates element-wise on ndarrays, e.g., np.sqrt(arr), np.exp(arr).

Question 36 Python for Data Science

How do you create a histogram in matplotlib?

pd.merge(df1, df2, on='key', how='outer') keeps all rows from both DataFrames, filling NaN where keys don't match. It measures the wall time and CPU time for the entire cell's execution (single run, unlike %timeit). plt.scatter(x, y) plots individual data points as dots on the chart. plt.hist(data, bins=20) displays the frequency distribution of data using 20 bins.

Question 37 Python for Data Science

How do you detect duplicate rows in a Pandas DataFrame?

np.eye(3) returns a 3×3 array with ones on the diagonal and zeros elsewhere. df.sort_values(['col1', 'col2'], ascending=[True, False]) sorts by col1 ascending, then col2 descending. sns.heatmap(df.corr(), annot=True, cmap='coolwarm') displays a color-coded correlation matrix. df.duplicated() returns a boolean Series that is True for duplicate rows.

Question 38 Python for Data Science

How do you convert a string column to datetime in Pandas?

pd.to_datetime(df['date_col']) parses strings into datetime64 objects. pd.DataFrame({'col1': [1, 2], 'col2': [3, 4]}) creates a DataFrame with two columns. Broadcasting is NumPy's ability to perform element-wise operations on arrays of different shapes by automatically expanding the smaller array to match dimensions. df['col'].apply(lambda x: x ** 2) applies the lambda function to each element of the column.

Question 39 Python for Data Science

How do you convert a column's data type to integer in Pandas?

df.dropna(how='all') removes rows where every column value is NaN. df[df['age'] > 30] returns a filtered DataFrame using boolean indexing. A tuple (rows, columns) indicating the dimensions of the DataFrame. df['col'] = df['col'].astype(int) casts the column to integer type.

Question 40 Python for Data Science

How do you generate 5 evenly spaced values between 0 and 1 in NumPy?

df['col'].interpolate(method='linear') fills NaN values using linear interpolation between existing data points. df['col'].fillna(df['col'].mean()) replaces NaN values with the column's mean. It returns True if value is NaN, None, or NaT (Not a Time). np.linspace(0, 1, 5) returns array([0., 0.25, 0.5, 0.75, 1.]).

Question 41 Python for Data Science

How do you create a line plot with matplotlib?

plt.plot(x, y) followed by plt.show() draws a basic line plot. df.groupby('category')['value'].mean() computes the mean of 'value' for each unique category. np.zeros(5) returns array([0., 0., 0., 0., 0.]). df.groupby('col').apply(my_function) applies my_function to each group DataFrame.

Question 42 Python for Data Science

What is the advantage of the categorical data type in Pandas?

df['col'].fillna(method='ffill') performs forward fill, propagating the last valid value. It reduces memory usage and speeds up operations like groupby on columns with a limited number of unique values. df['col'].unique() returns an array of all distinct values in the column. df['col'].nunique() returns the count of distinct values, excluding NaN by default.

Question 43 Python for Data Science

How do you write a DataFrame to a CSV file?

df.to_csv('output.csv', index=False) saves the DataFrame without the row index. arr > 5 returns a boolean array where each element indicates whether the condition is True or False. df['date_col'].dt.dayofweek returns integers 0 (Monday) through 6 (Sunday). df['col'].apply(lambda x: x ** 2) applies the lambda function to each element of the column.

Question 44 Python for Data Science

How do you create a scatter plot with matplotlib?

plt.scatter(x, y) plots individual data points as dots on the chart. Use plt.title('Title'), plt.xlabel('X'), and plt.ylabel('Y') before calling plt.show(). arr * 3 multiplies every element of arr by 3 without an explicit loop — the scalar is broadcast across the array. df['date_col'].dt.dayofweek returns integers 0 (Monday) through 6 (Sunday).

Question 45 Python for Data Science

How do you count the size of each group?

np.linspace(0, 1, 5) returns array([0., 0.25, 0.5, 0.75, 1.]). df['col'].fillna(method='ffill') performs forward fill, propagating the last valid value. plt.bar(categories, values) creates a vertical bar chart for the given categories and values. df.groupby('col').size() returns a Series with the number of rows in each group.

Question 46 Python for Data Science

How do you create a Pandas DataFrame from a dictionary?

df.sort_values('price', ascending=False) returns the DataFrame sorted from highest to lowest price. Feature scaling normalizes the range of features so that algorithms sensitive to magnitude (e.g., KNN, SVM) perform correctly. Common methods: min-max and standardization. pd.DataFrame({'col1': [1, 2], 'col2': [3, 4]}) creates a DataFrame with two columns. plt.savefig('plot.png', dpi=300, bbox_inches='tight') saves the current figure as a PNG file.

Question 47 Python for Data Science

How do you slice rows 1 through 3 (inclusive) from a NumPy array?

It selects all rows of the first column, returning a 1D array. pd.concat([df1, df2], axis=1) joins DataFrames horizontally, aligning on the index. Shift + Enter executes the current cell and advances to the next cell. arr[1:4] selects rows at indices 1, 2, and 3 (the stop index is exclusive).

Question 48 Python for Data Science

How do you rename columns in a Pandas DataFrame?

df.rename(columns={'old_name': 'new_name'}) returns a DataFrame with the specified columns renamed. plt.hist(data, bins=20) displays the frequency distribution of data using 20 bins. df['col'].unique() returns an array of all distinct values in the column. df.sort_values(['col1', 'col2'], ascending=[True, False]) sorts by col1 ascending, then col2 descending.

Question 49 Python for Data Science

How do you group a DataFrame by a column and compute the mean?

df.groupby('category')['value'].mean() computes the mean of 'value' for each unique category. df['col'].apply(lambda x: x ** 2) applies the lambda function to each element of the column. np.zeros(5) returns array([0., 0., 0., 0., 0.]). df[(df['age'] > 25) & (df['city'] == 'NYC')] — wrap each condition in parentheses and use & (and) or | (or).

Question 50 Python for Data Science

What does `arr[:, 0]` select from a 2D NumPy array?

df.iloc[:5, :2] uses integer-based indexing to slice rows and columns. fig, axes = plt.subplots(1, 2, figsize=(10, 4)) creates a figure with 1 row and 2 columns of axes. It selects all rows of the first column, returning a 1D array. np.sqrt(arr) applies the square root to each element and returns a new array.