(a) Explain different methods of visualising random forests. 
(b) Explain the use of the bivariate histogram and kernel density estimation with the help of a suitable
example.
(c) Explain the Trellis Paradigm. Why is it needed?
(d) Explain various regression visualisations with the help of an example

Write a Python program with suitable comments using pandas that performs the following tasks after
reading the file.
 identify missing values,
 remove duplicate records,
 replace inconsistent values,
 fill missing salary values using an appropriate method, and
 rename the column headings.
(d) Explain different techniques used for handling missing data. Discuss situations where each technique is
appropriate. (3 Marks)
8
(e) A dataset contains the ages of 500 customers ranging from 18 to 80 years. Explain different ways to
classify these customers into age groups using discretisation (binning). (4 Marks)
(f) Write Python code to detect outliers using the interquartile range method. Also, explain the strategies for
handling outliers in data with the help of suitable examples. (4 Marks)
Q2:
(a) Two data files contain customer details and purchase details having the same format. These files may
contain overlapped data. Write and explain the Python program to merge the two files. Make and state
assumptions, if any (3 Marks)
(b) Explain hierarchical indexing. Illustrate how hierarchical indexing can be used for reshaping datasets
with suitable examples. (3 Marks)
(c) A supermarket dataset contains the following fields: (3 Marks)
 Region
 Product Category
 Sales
 Profit
Write a Python program to perform the following tasks:
 group data by Region,
 calculate total sales,
 calculate average profit.
(d) Why is the need to create groups during data analysis? Explain with the help of an example. Explain the
working of the groupby() function in pandas. Discuss grouping using dictionaries, functions and Series
in Python with suitable examples. 
(e) Make a synthetic dataset containing monthly sales of three different products for one year. Write a
Python program using Matplotlib and/or Seaborn to generate a line plot and a scatter plot. Make suitable
assumptions. 
(f) Write a Python code to demonstrate the purpose of figures, axes, labels, legends, annotations and subplot
arrangements with suitable examples.
 

 A retail company has collected customer transaction data containing the following attributes: (3 Marks)
 Customer ID
 Age
 Gender
 City
 Purchase Amount
 Payment Mode
Describe how you would perform data profiling on this dataset. What are the summary statistics that you
would perform on this data?

Explain data wrangling. Discuss the importance of data semantics, set-based profiling, summary
statistics, and data profiling in preparing raw data for data analysis. Illustrate your answer with a suitable
example

Make a CSV file that contains employee information. The file should have a few missing salary values, a
few duplicate employee records, a few inconsistent department names, and some incorrect city names.

3:
(a) What is a scatter plot matrix? Why is it needed? Write a Python program to make a scatter plot matrix.
 (3 Marks)
(b) A weather department has recorded daily rainfall and temperature data for five years. Suggest suitable
graphical techniques for visualising this dataset. Justify your choices with reference to time series,
scatterplot matrices, and geographical maps. 
9
(c) Using Python, create a dashboard consisting of multiple plots arranged in a single figure. Demonstrate
how titles, legends, annotations and customised plot parameters improve the readability of the
visualisation.
(d) Explain any four applications of graphs in data science with suitable examples. (4 Marks)
(e) A social networking website wants to analyse the relationships among users. Describe how graph
representations such as bipartite graphs, hierarchical trees and graph drawing techniques can be used for
analysing the network. Illustrate your answer with suitable diagrams.