Exploratory Data Analysis (EDA)

Question 1

What is Exploratory Data Analysis (EDA)?

Answer 1

A process of analyzing and summarizing datasets to uncover patterns, detect anomalies, and check assumptions before modeling.

EDA helps understand the structure and quality of data.

Question 2

Which library is most commonly used for EDA in Python?

Answer 2

Pandas

Pandas provides powerful tools for data manipulation and summary statistics.

Question 3

What function in pandas displays the first few rows of a dataset?

Answer 3

df.head()

Question 4

What function is used to check the number of rows and columns in a dataset?

Answer 4

df.shape

Returns a tuple (rows, columns).

Question 5

What method provides an overview of the dataset, including column data types and missing values?

Answer 5

df.info()

Useful for checking null values and data types.

Question 6

Fill in the blank:

The method to calculate summary statistics for numerical columns is ___.

Answer 6

df.describe()

This function provides statistics like mean, median, min, max, and quartiles.

Question 7

Which pandas function counts unique values in a categorical column?

Answer 7

df[“column”].nunique()

The .nunique() method tells how many unique values exist.

Question 8

What method checks for missing values in a dataset?

Answer 8

df.isnull().sum()

It returns the count of missing values per column.

Question 9

True or False:

The .dropna() method removes all rows with missing values.

Answer 9

TRUE

Be cautious when using .dropna(), as it may remove important data.

Question 10

How do you replace missing values in a column with the mean?

Answer 10

df[“column”].fillna(df[“column”].mean(), inplace=True)

This fills missing values with the column’s mean.

Question 11

What function in pandas detects duplicate rows?

Answer 11

df.duplicated()

Returns a Boolean series indicating duplicate rows.

Question 12

Which visualization is used to check the distribution of a numerical column?

Answer 12

Histogram

Histograms show the frequency distribution of numerical values.

Question 13

What is the purpose of a boxplot in EDA?

Answer 13

To visualize the spread of data and detect outliers.

Boxplots display quartiles and outliers.

Question 14

Which measure of central tendency is most affected by outliers?

Answer 14

Mean

The mean is pulled in the direction of extreme values.

Question 15

What statistical measure is used to detect skewness?

Answer 15

Skewness coefficient

A skewness value > 0 indicates right-skewed data; < 0 indicates left-skewed data.

Question 16

True or False:

A correlation value of 0 means two variables are unrelated.

Answer 16

TRUE

A correlation of 0 indicates no linear relationship, but they may still be related in other ways.

Question 17

What visualization is commonly used to display correlations?

Answer 17

Heatmap

A heatmap visualizes correlation coefficients using colors.

Question 18

What does a scatter plot show?

Answer 18

The relationship between two numerical variables.

Useful for identifying trends and correlations.

Question 19

Fill in the blank:

A categorical variable is best visualized using a ___.

Answer 19

Bar plot

A bar plot shows the count or proportion of categories.

Question 20

How do you create a scatter plot using Seaborn?

Answer 20

sns.scatterplot(x=”col1”, y=”col2”, data=df)

Scatter plots help visualize trends between two numerical variables.

Question 21

Which method groups data by a categorical column?

Answer 21

df.groupby(“column”)

Allows aggregation of data by categories.

Question 22

What is the difference between a histogram and a bar plot?

Answer 22

A histogram is for numerical data, while a bar plot is for **categorical **data.

Histograms group data into bins; bar plots show distinct categories.

Question 23

True or False:

Outliers should always be removed.

Answer 23

FALSE

Outliers should be analyzed before deciding to remove them.

Question 24

What statistical method detects outliers based on quartiles?

Answer 24

Interquartile Range (IQR)

Outliers are values beyond Q1 - 1.5IQR or Q3 + 1.5IQR.

Question 25

What is the **range** of correlation values?

Answer 25

**-1 to 1** ## Footnote A correlation of -1 is a perfect negative relationship, while 1 is a perfect positive relationship.

Question 26

Which type of chart is best for showing **time-series** data?

Answer 26

**Line plot** ## Footnote Line plots show trends over time.

Question 27

# Fill in the blank: A __ is a technique used to reduce the number of features while preserving information.

Answer 27

Principal Component Analysis (PCA) ## Footnote PCA helps reduce dimensionality and improve model efficiency.

Question 28

Which function in pandas returns the **most frequent** values in a column?

Answer 28

df["column"].value_counts() ## Footnote Useful for analyzing categorical variables.

Question 29

What is the purpose of *feature engineering* in EDA?

Answer 29

To create new meaningful features from raw data. ## Footnote Helps improve model performance.

Question 30

Which function converts **categorical** variables into **numerical** format?

Answer 30

pd.get_dummies() ## Footnote Creates one-hot encoded variables.

Question 31

What is **multicollinearity**?

Answer 31

When two or more independent variables are highly correlated. ## Footnote Multicollinearity can distort regression models.

Question 32

Which statistical test is used to check **normality**?

Answer 32

Shapiro-Wilk test ## Footnote A p-value < 0.05 suggests non-normal data.

Question 33

What does *kurtosis* measure?

Answer 33

The "**tailedness**" of a distribution. ## Footnote High kurtosis = heavy tails, low kurtosis = light tails.

Question 34

How do you detect **missing** patterns in data?

Answer 34

Using a missing value heatmap **(sns.heatmap(df.isnull(), cmap="viridis"))**. ## Footnote Helps visualize missing data structure.

Question 35

What is a **QQ** plot used for?

Answer 35

Checking if data follows a **normal** distribution. ## Footnote A straight line indicates normality.

Question 36

# True or False: **Normalization** scales data to have a mean of 0 and standard deviation of 1.

Answer 36

**FALSE** ## Footnote Normalization scales data between 0 and 1; standardization sets mean = 0, std = 1.

Question 37

What is **data leakage** in EDA?

Answer 37

When training data contains **future information** that shouldn't be available. ## Footnote Can cause unrealistic model performance.

Question 38

What does a **pairplot** show?

Answer 38

Pairwise relationships between multiple numerical variables. ## Footnote Created using sns.pairplot(df).

Question 39

What is the purpose of **dimensionality reduction**?

Answer 39

To **reduce** the number of variables while retaining information. ## Footnote Helps improve computation efficiency.

Question 40

# True or False: **Log transformation** can help with skewed data.

Answer 40

**TRUE** ## Footnote Log transformations reduce right skewness.