PRAVEEN S

2214100868
BACHELOR OF BUSINESS ADMINISTRATION (BBA)
SEMESTER - 5
COURSE CODE & NAME DBB3102-BUSINESS ANALYTICS
 SET 1

1.
(a) The three main categories of business analytics are:(a) The three main categories of
business analytics are:

- Descriptive Analytics: This involves looking at the data, to try and see what has occurred in the
past, and the reasons why it occurred. Businesses record their key performance indicators and
various other measures with a view to identifying patterns. An example entails; sales reports,
web traffic, social media analytics among others.

- Predictive Analytics: This involves the use of an algorithm to predict future events based on
the past occurrences of similar events. Examples include sales forecast, customer attrition rate,
tourists’ arrivals, among others The main advantage is that of being in a position to act on data
generated by the model.

- Prescriptive Analytics: This draws out decision possibilities and advises on the best course of
action to be taken in order to pursue a given goal. It uses optimisation algorithms, simulation
models, decision sciences and AI to offer the right path forward. Some of them are customer
acquisition cost management, demand forecasting, dynamic pricing strategies, inventory
management, targeted marketing, etc.

(b) Some examples of how business analytics can transform the tourism industry include:(b)
Some examples of how business analytics can transform the tourism industry include:

- Predicting tourist arrivals to various destinations during holiday and non-holiday seasons using
time series models. This ensures right capacity and workforce planning is done to ensure that
the organization meets its demands efficiently.

- Most often computational models to forecast specific cancellation rates, no-shows etc. using
historical data. This makes it possible to overbook to desired levels aiming at maximizing on the
revenues.

- Female tourist’s pain points’ discovery on social media using sentiment analysis to enhance
service delivery. One of the applications where natural language processing can be used is text
mining of unstructured textual data.

- Tourist guides to give recommendations on what tourists can do, places they can visit, where
to get food and other services they may require at certain places and at certain times. This
improves experience.
- Depleting hotel room inventory and airline seat inventory in an attempt to maximize the
revenue considering the demand forecast. Prices can be changed very frequently as the
forecast changes and this is a unique advantage of self-generated forecast.

- Staff scheduling and rostering to model the best approach to allocate staff to the tasks and
match the supply with the demand. It also eliminates long hours of waiting for the doctor as well
as cutting down the expenses of hiring more employees.

2. Information management is important as it helps in organizing the data in order to gain

storage, retrieval, analyzing, and using the information. Well-organized data brings several key
benefits:Well-organized data brings several key benefits:

- Faster search and retrieval: The systematic arrangement of information through

implementation of categorization, metadata, folder structures, central databases, etc. prove
beneficial while trying to find the precise data required in the future. Failure to organize data
leads to complications, notably, the difficulty of sifting through heaps of unstructured masses of
data.

- More insights from analytics: Information needs to be arranged in standard and consistent
formats which is much easier in terms of aggregation, processing and analysis of the same. This
way, rather than finding endless fragments of information, data analysts can reveal the most
crucial patterns and trends with the data neatly picked and put together. Completeness and
consistency of the data determine what can be analyzed.

- Reduced storage needs: These measures ensure that data is not stored in two or more
locations to prevent cases of duplication in the storage space. Minimizing data duplication
through creation of master databases is resourceful as it saves on space that would have been
used by duplicate data. Another way that compact storage is achieved is through organizing
files and documents in folders, archiving, and purging data that is no longer relevant.

- Improved data integrity: Business principles such as creating single source of information, data
scrubbing, data dictionary creation and standardization, data auditing, controlling and security
are particularly important in maintaining the integrity and consistency of data over a period of
time. Every system that receives data from another system is protected from the possibility of
errors being passed to it.

- Support for governance/compliance: Any organization that comes under legal requirements
that pertain to data management, retention, privacy, and so on, can easily establish controls
when the information assets are categorized. Strong parameterized databases, authorization
mechanisms and best practices such as data management and disposal help in compliance.

It is a well-known fact that demanding time and planning initially will result in great benefits after
in the form of insightful analytics, malleability of information for various functions, effective
utilization of storage and systems, cost and risk minimization regarding low quality data. With
the volumes and types of data remaining high and further increasing, the adoption of an
intelligent approach to data organization is one of the most valuable endeavors any modern
organization can undertake.

3. Data transformation is the process of converting data from its original raw format into a more
useful and efficient format for various purposes like analysis, reporting, visualization etc. Some
common data transformation techniques are:Data transformation is the process of converting
data from its original raw format into a more useful and efficient format for various purposes like
analysis, reporting, visualization etc. Some common data transformation techniques are:

- Data Cleaning – It means identifying and purging errors or inconsistencies found in data to
enhance its quality. The following methods were employed; Outlier removing which involved
identifying and eradicating them from the databases, Missing values which were addressed
through imputation methods, Duplicate records which were solved by treating the duplicates and
Data standardization which involved ensuring that the information was in the right format.

- Data Integration – It refers to integration of data from different sources in a way that data about
the same client have the same identification number for easy comparison.
Techniques used include schema matching where different match point of two tables is
identified, entity consolidation, data warehousing where data is collected from different sources
for analysis and for transformation processes such as ETL (Extract, Transform and Load) etc.

- Data Reduction – It is concerned with the raw data and its objective is to condense the data so
that analysis can be performed rapidly while keeping the trends and patterns intact. Some of the
techniques used are aggregation which involves grouping of similar data, data cube
constructions, dimensionality reduction which reduces the dimensions of data, data
compression which involves compressing data, numerosity reduction which involves sampling,
discretion and concept hierarchies.

- Data Discretization – It means transforming data into discrete terms from continuous terms to
enhance easy analysis. Examples of the techniques used include equal width binning, equal
frequency binning, entropy-based discretization.

- Data Normalization – It transforms the range and scale of the attributes data such that all the
attributes are transformed to range between 0 and 1 to reduce the impact of any large scale
attribute. Some of the techniques that are usually employed include min-max normalization,
z-score standardization, decimal scaling amongst others.

- Data Masquerading – It seeks to alter data values or format in order to obscure identity or
eliminate sensitive information that may be deemed unfit for sharing in the commercial market.
Some of the methods implemented include:k-anonymization, l-diversity, data embellishment,
data scrambling, data encryption and others.
In other words , data transformation involves data cleansing, data integration, data reduction,
data discretization, data normalization and data generalization in order to transform the raw data
into a simpler type of data which is easier to analyze and apply for Business Intelligence and
data privacy. The information obtained from the transformation leads to knowledge that is useful
in making crucial business decisions. Selecting proper techniques is contingent on specific use
case needs at the same time for preventing overcomplicated techniques which might require
lots of resources and time.
 SET 2

1. (a) Business analytics is described as the techniques or approaches, tools or technologies

that the business organizations employ when doing analyses on the business data that has
been collected to acquire useful information that can be used enhance the performances of the
business. They extend over a broad scope of applications to help the firms to acquire the data,
document it, and analyze it to have knowledge of the some specific pattern, trend, and
information for the strategical and taktical business management decisions. The main
advantage of OD is in increasing the productivity of organizations.

(b) Business analytics can improve overall organizational efficiency in several ways:Thus, the
business analytics can enhance the organizational effectiveness in the following ways:

- Analyzing business – since figures concerning sales, stock, supply chain, production and so
on are measured organisations may choose aspects that requires improvement. Analytics is
also about applying other consideration like sensitivity analyses to predict effects of certain
decisions with regard to resources, capacities, quality and other factors which in essence
means efficiency of various processes, more rationally made decisions, less operation costs and
high profitability.

- Improved sales and operations planning – In the use of the ‘predictive model’, it can be seen
that sales and operations could be easily predicted in the future. This brings a notion that it is
possible for organizations to control the level of flexibility of the supply chain so that adequate
stock and products that would support the expected demand are produced. When forecasting is
done there is less wastage h and unsold products, costumers are satisfied and attended to with
lesser cost.

- Notifications – Instead of letting human analyzer go through the data and come up with the
results, it can integrate and use aspects such as machine learning, artificial intelligence that can
define patterns, irregularities, and relations that even the analyzer would not. Such ideas help
the organizations to carry on improving the operations and effectiveness of the business.
Self-organizing insights also enhance the efficiency as well as effectiveness of decision-making.

- Further directly busines-customer interaction – Customer analytics helps organizations to sort

out the customers based on buying behavior and their inclinations. This could involve individual
offers and promotions on products and services, individual tagging of prices and ways and
means of services to the customers. The addition of cross selling and upselling functions aimed
at the person enhances the probability of lead capture on the webpage, and total sale. It also
directly impacts customer satisfaction and customer loyalty since the customers are more likely
to visit places were they encounter friendly people.

2.
a) Importance of data quality and best practices for maintaining data quality:importance of data
quality and best practices for maintaining data quality:
Impact: This is an area of concern to any organization that deals with large amounts of data for
decision making, analysis, reporting and other business purposes. Accuracy and reliability of
data that has been collected together with its comprehensiveness and updating is very useful
when one wants to get the right insight or make the right decision. Poor quality data can
negatively impact operations and strategic initiatives in numerous ways:
- If information is incomplete, untimely, irrelevant or inaccurate conclusions and decisions arising
therefrom result in the organization being placed at a position where is financially ‘lose/losses’,
lack resource, or have unsatisfied customer.
- The absence of total information negates one’s ability to cover the round as a way of scouting
for future use in planning specially for analytical purposes. Some important data may be shed.
- Some specific challenges are due to the problem of data incoherencies essential for
aggregation and analysis. This may mean it has contradicting info which is rather inconclusive
Its data may contain contradicting info and this can be inconclusive if various sources we
interact with during research contain such data.
- Consistent outdated data creates a wrong perception of the existing reality which is very
undesirable in risk management. The idea of working on data that would be, perhaps,
non-existent within days exposes the emotional and ad hoc approach to running a business.
(b) Best practices for maintaining high data quality include:
- Adopt data stewardship as a way of maintaining compliance on the proper policies, standards,
and authority on operational data as an enterprise resource.
- What is more, when constructing profiles for the data sets, it is possible to identify the quality
issues, duplication, or other issues regarding integrity that should be resolved. Establish metrics
and thresholds.
- Carry out processes like general record cleaning which may include normalization and
deduplication of records, initial record error correction. Automate wherever possible.
- Temporarily store information about data in a process and gather metadata about data starting
from source systems up to the transformations and usage stages. The concept behind this step
is in an attempt to seek the link or relation of data elements that have been traced so as to
identify its source.
- Subordinate reference models as well as information, master data and data repository and
description.
- Use automatic rule-based validation every time data is intake and periodically review the data
quality problems for correction.
- Ensure that the teams involved in data management are provided for and trained to ensure
that they maintain quality while doing their work.

3.Summary of the key steps involved in the data cleaning process:summary of the key steps
involved in the data cleaning process:
Data cleaning is the process of locating and correcting data anomalies which will improve the
quality of data. It typically involves the following key steps:The process is usually undertaken in
the following process elements:
i) Data quality assessment – Further analysis will be made to the collected data and this
process will entail identifying some of the issues that are with the data for instance, missing
values, duplicated data among others. This tends to give an indication of the level of cleaning
that may be necessary and indeed if the cleaning of the area is work that can be easily
accomplished.
ii) Indicate a cleaning process – In this case, as it is noted that different types of issues are
identified, the cleaning process is selected based, correspondingly. This can consist of issues
such as rectifying structural flaws, eradicating undesirable information, handling cases of
missing values, and overcoming disparities among other others This is because the techniques
utilized are inclined by the measures of data tidiness that might be needed as well as the
objectives that are likely to be achieved.
iii) Data Transformation– This is in regards to the format, structure or even the values of data
that may have to be transformed and altered before the use in individual organizational
processes from data that has been delivered from various sources. Operations like data
consolidation, normalization and integration are some of the methods which is used in
reformatting of the raw dataset.
iv) Other tests like business rules, correlation, duplication and referential integrity check etc are
carried out with a view to establishing further invalidations. Any action that is contrary to the
expected norms is documented down.
v) Compensating errors – This means that the errors need to be adjusted manually; the dataset
is physically Go through to correct the errors strictly in accordance with the outlined validation
rules and going through any form of reasonability.
vi) Anomalies – These are managed statistically and may also be prescreened for authenticity
which enable the analyst to deliberate on their nature and thus determine whether they should
be expunged or have their values modified in some way.
vii) Completeness convergence- After the data cleaning process, it is examined if the required
data is sufficiently available for the intended application. There might be some remaining
missing values that are essential to analyze, then, they are imputed or otherwise mentioned.
viii) Record findings and actions – All data quality issues that have been identified and the
detailed processes of data cleaning performed, changes that have been made, and all other
notes that is taken such that these could be used in future data extracts about the lessons learnt
and the processes involved.