1. BODY OF THE COURSE
COURSE NAME – BUSINESS ANALYTICS
ONLINE PLATFORM – UPGRADE
NUMBER OF WEEKS – 6 WEEKS
ABOUT – 01. INTRODUCTION.
02. EXPLORATORY DATA ANALYSIS
03. BIG DATA
04. MACHINE LEARNING
05. DEEP LEARNING, NEURAL NETWORKING AND
NATURAL LANGUAGE PROCESSING
06. APPROACHES TO SOLVE PROBLEM
07. LEARNING OUTCOMES
08. CONCLUSION
2. LEARNING EXPERIENCE
INTRODUCTION
Business analytics – It is the process by which businesses use statistical
methods and technologies for analysing historical data in order to gain new
insight and improve strategic decision-making.
Data analysis v/s data science
- They are 2 sides of a coin. Analytics is the discovery, interpretation and
communication of meaningful pattern in the data whereas Data science
also means extracting insights from data and help in data driven decision.
Earlier data was not so big, knowledge of statistics was good enough to
analyses these data. This is was era of analytics. Overtime, data set has
become more complex and managing them require specialized
engineering skill sets. So, we require a strong hold on statistics. Therefore,
this is an era of data science.
What kind of data is captured?
First is transaction history like User id, data, time, cart value, delivery date,
return tag and many more. Next is item meta tag like items ID, name, category,
MRP and Selling price, Procurement date etc. third kind of data is Supply chain
history here they capture Order ID, Item ID, Inventory location delivery – date
and time, Pin code and executive. Fourthly User meta tag it capture – User id,
Email id, gender, date of birth, etc. lastly vendor details – Vendor’s ID, location,
rating category and much more. This are some of the data the e-commerce site
capture.
EXPLORATORY DATA ANALYSIS (EDA)
- It is one of the important steps in any kind of data analysis. It refers to the
critical process of performing initial investigations on data so as to
discover patterns, to spot anomalies, to test a hypothesis and to check
3. assumptions with the help of summary statistics and graphical
representations.
- Once the data is explored and have some findings, it is imperative to be
able to present it in a format that can be understood by the senior
management. This is where reporting comes into the picture. To do
further analysis cleaning the unnecessary data is must. Data
cleaning, feature engineering all fall under the broad category of data
preparation. Once the data is prepared, univariate data analysis can be
done on it.
BIG DATA
- For example, The Amazons and the Flipkarts of the world have an
extremely huge amount of data. So huge that it becomes difficult to
analyse it on a single computer. We need a whole different infrastructure
to deal with it. This huge amount of data is termed as 'big data' and
analysing it is termed as big data analytics. It is characterised by 3 Vs -
Volume, Velocity and Variety. Volume refers to the size of the data,
velocity refers to the rate at which the data is being received, and variety
refers to the different types of data that we may get - images, text,
numbers, speech, videos etc.
- Data Architecture: Data architecture is a set of rules, policies, standards
and models that govern and define the type of data collected and how it is
used, stored, managed and integrated within an organization and its
database systems.
- Parallel Computing: Parallel computing is a type of computing in which
many calculations or the execution of processes are carried out
simultaneously. Large problems can often be divided into smaller ones,
which can then be solved at the same time.
MACHINE LEARNING
4. There are 2 types of machine learning i.e., Supervised Learning-it is a type
of machine learning algorithm in which a system is taught to classify input
into specific, known classes. Classification is one such technique which
classifies data points into one of the various possible classes. And next is
Unsupervised Learning where a class of machine learning algorithms
designed to identify groupings of data without knowing in advance what
the groups will be.
DEEP LEARNING, NEURAL NETWORKING AND NATURAL
LANGUAGE PROCESSING
- Deep Learning- Typically, a multi-level algorithm that gradually identifies
things at higher levels of abstraction. For example, the first level may
identify certain lines, then the next level identifies combinations of lines as
shapes, and then the next level identifies combinations of shapes as
specific objects. As we might guess from this example, deep learning is
popular for image classification.
- Neural Networks: A robust function that takes an arbitrary set of inputs
and fits it to an arbitrary set of outputs that are binary. In practice, Neural
Networks are used in deep learning research to match images to picture.
- Natural Language Processing: A branch of computer science for parsing
text of spoken languages (for example, English or Mandarin) to convert it
to structured data that we can use to drive program logically.
- Artificial Intelligence is essentially teaching a machine to think like a
human. It is surprisingly difficult. Suppose we are watching a cricket
match. We can look at the eyes of the batsman and know what shot he
will play. Now if we can train a machine to predict the same, we can
imagine what a big breakthrough that is. That is Artificial Intelligence.
5. APPROACHES TO SOLVE PROBLEM
1. UNDERSTANDING THE BUSINESS PROBLEM
- To start with any analysis, we should first understand the problems, which
the business is facing. After doing plenty of self-research, we need to
understand the problem from the ones who are facing it. Here comes the
part that we term ‘interviewing’. To understand a problem completely, we
will always need to interact with multiple people in the company.
Interviewing people all the time to gather information. Different job roles
will require to interact with different sets of individuals, but the task of
interviewing will remain the same across all of them. Interviewing is an
important segment, while interviewing following things should be kept in
mind – Turn off all the distractions (mobiles, laptops, etc.) around us, use
pen and paper to prepare notes, be patient Don’t be anxious to reply,
Pause Think and then Ask, Playback understanding with the interviewee
for their validation.
Frame works - To overcome this issue, people have developed specific
patterns of asking questions over the years, which we call ‘frameworks.’
There are multiple frameworks available at our disposal, and we need to
pick the one that is the most suitable for our case. There are 3 important
frameworks- 5 WHYs, 5 HOW’s, So what? And 5 W’s. All three
frameworks are useful to understand the context of the problem. It is
helpful in identifying the root cause for a problem. We cannot cover all the
domains using these frameworks, but they give a sense of the problem.
- The SPIN framework (Situation, problem, implication, need-payoff
questions) starts with asking about the current situation and helps us to
visualise the entire journey, from when the problem arises to what will
happen when the problem is solved. It is an excellent approach to follow,
6. as it helps the client (internal or external) realise the extent of both the
problem and the solution.
2. FORMULATING HYPOTHESIS
After understanding the dept of the problems we should ask the questions in an
interview to get the insights regarding the problems the firm is facing. When we
understand the problem, we need to explore the reasons that may be behind it.
After the interviewing process if we think that we have identified the problem,
we are mistaken. It is just one possible reason for the problem that the
company or client is facing. There can be multiple reasons for the problem.
Also, even if we believe that we found the correct reason, first we need to verify
whether there is any data that supports it and then employ the resources to
solve the issue. This test is required because if there was a different reason
behind the problem, all the efforts would then go in vain. Therefore, we need to
realise two things:
1. When we are exploring possible reasons for any problem, we need to cover
all the aspects of it. For this, we will again come across various frameworks that
can be applied to follow a structured approach.
2. Also, the possible reason that we discovered is what we call a hypothesis. A
hypothesis is a possible explanation which has been prepared based on limited
evidence and needs to be validated by further investigation.
BUSINESS MODEL CANVAS
Business Model Canvas is a strategic management template for developing new
or documenting existing business models. It helps you cover all the aspects on a
sheet of paper in a structured format. It is distributed under a 'Creative
Commons license' from Strategyzer and can be used without any restrictions for
modelling businesses. This model is adaptive to all business and it covers all the
domain of the business in a simple and efficient manner.
FRAMEWORKS
After the interviews, we will have possible reasons or causes for the problem
that the company or the client was facing. The focus point is that it is still a
7. "possible" cause. Therefore, we use the term 'hypothesis'. Now, we will learn
the process of formulating the hypotheses using multiple frameworks. Some of
frameworks are:
1. Issue Frame Work - Issue tree framework is one of the most effective
methods to approach a problem. It works by disintegrating the problem into
sub-components. The big complex problem is continuously decomposed into
simpler issues. At last, we will end up with a bunch of hypotheses.
2. Specialised frameworks –The issue tree framework can be clubbed with
different frameworks to provide a complete view of the problem. Based on the
domain of the problem that we want to solve, there are multiple frameworks
available to apply. For example, if the company is trying to create awareness in
the market, it is a marketing problem. If the company wants to optimise the
process of manufacturing, it will be an operations problem. These cases are
different in nature and should be handled differently. In specialised frameworks
there are some of frameworks which are very effective and help us to approach
the problem in a structure manner:
o Business segmentation - This framework is useful when the company or
client has a spread over multiple businesses. Like in issue tree, you will be
required to break the entire company into the respective businesses in
which it operates. For example, e-commerce giants like Amazon, Flipkart,
etc. provide a wide range of products like clothing, electronics, etc. If the
company is facing any problems, you will analyse these segments
separately because the characteristics of each business can be different
from one another.
o Profitability Analysis- This is also an extension of the issue tree
framework. When you are trying to break a branch into further sub-
components, one way to segment is based on the profitability of the
products or services. Here, the main aim will be to analyse the high yield
products prior to others.
o SWOT Analysis - The SWOT analysis helps you segment the processes in
the company based on the capabilities and points out the critical concern
areas for the company.
8. o The balanced scorecard - The idea around the framework is that
your learning and growth will help you to handle the internal
processes better. The betterment of operations will result in a reduction in
the process costs and improve customer experience. Better customer
experience will drive the revenues upwards. Hence, the financial aspect of
the company will grow. You can analyse these four aspects of the
company and formulate hypotheses around them which can be tested in
the later phases.
o 4Ps Framework - The 4Ps framework or the marketing mix model is a very
powerful tool to check the marketing strategy of the company. It focuses
on the following Ps: Product, Price, Promotion, Place. It serves the purpose
of solving the problem for the customer.
o 5C’s Framework - The 5C framework is a very useful framework to
understand both, the internal and external environments in which the firm
is operating. It helps you to identify what is helping the company to
succeed and which factors are restricting it from achieving what is
expected.
These frameworks should help to reach the final hypothesis at the end of the
interviewing process.
3. COLLECT DATA
The next step that follows the hypothesis formulation is to collect data to
validate them. Once the validation of hypothesis is done, we can start working
on a solution around that root cause. Collect the data with regards to the issue.
The data should be collected for a period of 3 to 6 months on the reported
problem, as well as linked issues. This step is essential to validate the hypothesis.
Data can gather by reviewing the business plan, Interviewing the data expert,
analyse forms and reports etc.
There are various types of quality issues when it comes to data, and that’s why
data cleaning is one of the most time-consuming steps in data analysis. In real-
world scenarios, the data you need to analyse often come from a third party,
9. clients, etc., and the data collection/entry methods, etc. often lead to errors,
due to which cleaning the data becomes crucial.
4. ANALYSE DATA
Once we have obtained the data the next step would be to analyse it to observe
any patterns that validate or refutes the hypotheses that we’d made in our
earlier steps. In the business problem-solving procedure this part comes
immediately after the collection of data.
o PATTERNS OF INSIGHTS
Irrespective of the business problem that any industry is trying to solve,
the insight that is generated to solve that problem has some underlying
common patterns.
These patterns can be classified into five categories:
• Unknown Result: When the result is unknown and of significance
• Surprising Extreme: When the result is the highest or the lowest and
was not expected.
• Surprising Comparison: When two values are compared, and the
resulting inference is surprising.
• Significant Outliers: When a value is unusually large or unusually small.
• Abnormal Distribution: When a variable shows unusual trends.
Even though hundreds of results are generated from the given insights, the
corresponding insights need to follow two criteria to become insightful. These
criteria are: check if the insight is interesting and check if the insight is useful.
o DOCUMENTING INSIGHTS
Once the insights are generated, it’s crucial that we’re able to
communicate the results effectively to your audience. One of the most
prevalent ways in which consultants and analysts communicate insights,
the PYRAMID PRINCIPLE is crucial in conveying the message to the
audience very quickly and efficiently. This principle works well because of
10. two reasons: It’s concise and save the time. Pyramid principle can be
utilised in a variety of formats like slides, e-mails, etc. Therefore, we must
practise the skill of communicating the insights that we have using this
principle.
5. PRESENT FINDINGS
Storytelling is important because, research has shown that storytelling is more
persuasive, and people tend to remember stories far better than simple
statistics. The first element of storytelling that you need to learn is the concept
of using effective visualisation. visualisation is important because it makes the
available insights more digestible and easier to interpret. visualisation can be
possible for two types of variables: Qualitative and Quantitative. Visualisation of
Quantitative Variables – Scatter plots, Line charts, Histogram and for
Visualisation of Qualitative Variables – Pie chart, Bar chart, stacked bar chart are
used to analyse the data.
o VISUAL DESIGN PRINCIPLES - The next important aspect of storytelling is
the concept of visual design principles. This uses two key concepts.
1.Trade-off between Accuracy and Precision and 2. Drawing Attention with
Text and Visuals.
o STORY BOARDING - Storyboarding uses the concept of Pyramid Principle
to explain the main points and the supporting arguments and uses
additional nuances like placing information at the right place, removing
superfluous information and visually linking the items. It helps to
assimilate information, identify gaps in analysis, helps to avoid redundant
works.
11. LEARNING OUTCOMES
This course helped me to learn and understand how to analysis the basic
data. It will ultimately help to spot new business opportunities, cut costs,
or identify inefficient processes that need reengineering.
This course gave us a live case study and had a demonstration class to
understand the different types of data and extract root cause of the
business through interviewing, searching or collecting the data from the
clients. when we are interviewing to solve a problem, we will come across
people with different traits and the same approach will not work for all of
them. There are 4 types of interviewee – The old hand, The Weasel, The
Stone Face, I know it all. Tailoring the conversation according to their type
can result in good information source.
And then using different types of frameworks and formulating the
hypothesises and taking that hypothesis into consider to collect the
appropriate data.
Collection of data from right source is important. Once the validation of
hypothesis is done, we can start working on a solution around that root
cause. After collection of data is done, cleaning of data is an important
procedure. After collecting the data next procedure is to analysis it. And to
analysis the data EXCEL is the good platform
This course taught us different types of formulas, graphs, diagrams can be
used to analysis the data. This will help us to understand every corners
and aspects of the data. There is important tool called CONDITIONAL
FORMATING which can be used to understand the data more efficiently.
There is one more important tool PIVOT TABLE this tool converts the data
12. into table form and brings out the different form of information according
to our convenience.
Lastly, they thought us how to present the findings, we should first
choose the top insights, weave into a story, validate the outcome of the
message, and lastly bring out the impact. While
CONCLUSION
1. This course is an opportunity for me to understand about the
BUSINESS ANALYTICS
2. It gave me understanding about the process of the analyzing the
data.
3. It is good platform to do further studies.
4. It brings out all positive and negative aspects of the businesses.
5. This course helps us to know what skills and knowledge we have to
improve in coming time.
- Lastly, I would like to thank all the respected teachers and
commerce department for giving me this opportunity, it was
great learning experience to me while learning this course.