Introduction to Data Analysis

Course Outline

I. What are BI and DA?

      - Definitions of BI

      - History of BI

     -  How is BI used to help Businesses

     -  Definition of DA

     -  Relationship between BI and DA

 II. Data Here, There, and Everywhere!

       - Oracle study on business data preparedness

    -   Overview of Study Findings-overwhelmed by volume of data and inability to     utilize data effectively

        Possible solutions  to data overflow problems

III. Got Data? The Unique Role of the Data Analyst

        - Role of a Data Analyst

        - Skill set required to be an effective Data Analyst

Exercise: "Channeling  Your Inner Analyst"- Students are told to imagine receiving a memo from their supervisor explaining that the company is downsizing.  They are expected to take on additional responsibilities including doing data analysis. They must rewrite their current job description to include the new data analyst duties.

FACTS or Feelings: Your Choice

As data becomes more widely available, businesses are finding more success in adopting a fact based decision model rather than relying on traditional intuition alone. In this module, we examine more closely the two types of decision models businesses use as well as the benefits of the fact-based model. We cover the steps of the Rational Decision Model , a fact based method for decision making.

IV. Fact- Based Decision Making Process

     - The two types of Decision Models Businesses use

      - The Benefits of Fact- Based Decision Making

      -  Rational Decision Model : Six- Step Method

      - Pal's Diner: An Example of how the Rational Model is used in practice

Exercise: " Who's The Boss?"- Students are divided into groups, and told to imagine that they are the CEO of their  own company. They define a business related decision that they need to make and then apply the steps of the Rational Decision Model to arrive at the final conclusion.      

'BIG DATA' ANATOMY

In this module, we revisit the Big Data trend with a more detailed focus. We begin by defining the buzz word-"BIG DATA" , examining its core attributes, and outlining the factors that contribute to data being  'big'. We explore how businesses collect structured and unstructured data, and the challenges they face in storing and effectively using both types of data.

  V. Big Data Anatomy

      - The Attributes of Big Data

              - Definition of Big Data

              - The 4 V's of Big Data

             - Structured versus Unstructured Data

             - The Challenges of Big Data

Exercise: "Camp WoeBeData"- Students are asked to describe some of the big data challenges that their companies face and to outline what steps are being taken to address the problems.

GETTING TO KNOW YOUR DATA

In order to better understand how to analyze data, we first have to comprehend its depth. This  requires drilling deep beneath the server it is located on and understanding its composition. Assume we are given a structured data set with labeled columns  and completed rows . There are plenty of ways  to summarize the story behind the data, but we cannot dive in without first  getting to understand its fundamental structure. We begin by classifying the collected data as quantitative or qualitative. Then we further classify our column variables according to the way data is measured: nominal, ordinal, interval, or ratio. It is only after understanding this classification that we are able to proceed to the next step of choosing the appropriate analysis techniques which correspond  to nominal, ordinal, interval or ratio variables.

VI. Getting to Know Your Data

      - Data Types: Qualitative versus Quantitative

     - Taking a Closer Look: Data Measurement

              Four Types of Data Variables

                   Definition  and  examples  of Nominal Variables: Name only

                   Definition and examples  of  Ordinal Variables: Order Matters

                   Definition and examples of Interval Variables

                   Definition and examples of Ratio Variables

                   Summary of Statistics/Operations  that can be performed on each type

Exercise:  "Marketing to Low Renters" - Students  are told to put on their data analyst thinking caps. They have been employed as a junior data analyst for a Marketing Company whose goal is to make a marketing campaign for a client who plans on targeting the 'needy' population. Students are given a public housing data set and told to classify each variable according to its measurement.

DATA VISUALIZATION

A picture is worth a thousand words, and there definitely is no exception when it comes to summarizing data. This module is dedicated to highlighting the importance of visualizing data, and how the human eye depends on visual representation to get a quick sense of data relevance. Visual representation is the audience's first impression of the data and forms a crucial step in inviting and maintaining genuine interest in a subject matter. We demonstrate how to create colorful, easy to understand tables, charts, and graphs that aid in helping us convey the story behind the data set being analyzed.

VII. The Fundamental Ways we use data Visualization techniques

             -The five ways we use data visualization techniques

VIII. Displaying Tabular Data in Excel

             - How to create custom tables in Excel

             - How to Sort/Filter tabular data

              -How to create and manipulate pivot tables

IX. Using Charts and Graphs to Communicate Data

             - How to create Pie, Column, and Line charts using Excel

             - Communicating effectively using different chart types

              - How to choose the correct chart to display the correct data type

Exercise: "Table Mining"- Students develop tables to summarize trends in a data set related to Low Rent and Section 8 housing.

Exercise: "Charting Poverty"- Students develop charts and graphs to summarize the poor housing epidemic in a public housing data set.

NUMERICAL DATA SUMMARIES

Another way that data analysts summarize data is by providing a single number , or summary statistic, that has meaning. This module explores how the mean, median, and mode can be used to summarize the center of discrete and continuous grouped data . The range, standard deviation, and inter-quartile range measure the dispersion in the data set and provide information about how data points are spread.

X. Using Numerical Descriptives to Summarize Data

        - Measures of Centrality: Mean, Median, Mode

         -Format of Data Values: Grouped Discrete and  Grouped Continuous

         -Formulas for the Mean

                Examples: Applying 3M's to Grouped Discrete and Grouped Continuous Data

         -Measures of Spread: Standard Deviation, Range, Inter-quartile Range

                Examples: Applying Measures of Spread to Grouped Discrete and Grouped Continuous Data

Exercise: "Faulty Wear"- Students use mean, median, and mode to summarize information about returns from a department store.

Exercise:"Faulty Wear: Take 2"- Students use measures of spread to summarize the distribution of faulty garments from a department store.

QUANTIFYING UNCERTAINTY

Probability is not only the most important data analysis foundation, but it is by far the strongest measure that businesses can use to quantify uncertainty. There are risky decisions that businesses encounter when investing in certain stocks and taking a chance on whether its value will rise. The insurance industry calculates the most probable life expectancy for the population and bases its rates on that uncertainty. There are an abundance of examples that illustrate why understanding probability benefits the business industry, so this module is designed to expose students to a solid understanding of the topic. We use simple, easy to understand examples to introduce students to traditional and conditional probability. Then we follow up with other business  probability applications that involve relative frequency and expectation.

XI. Probability: Quantifying Uncertainty

         - Origin of Probability

          -Probability: Examples of Business Applications

          -Traditional definition of Probability

           -Simple Computation: The TopBottomFraction Method

           - How to calculate probabilities from contingency tables

           - How to Calculate conditional probability from contingency tables

            - Applying probability to calculate relative frequency

            - Applying probability to calculate expected value

             -Using Expected Value in Decision Making

Exercise:" Pocket Probability"-Students practice calculating basic probabilities using the change in their pocket.

Exercise:"Magazine Money"- A magazine subscription service conducted a survey to study the relationship between the number of subscriptions  per household and family income. Students use contingency tables to answer questions related to conditional probability.

Exercise: "Home Sweet Home"- Students consider an international data set involving the different types of home dwellings located in Bradfield, England. They use information from contingency tables to calculate regular and conditional probabilities.

Exercise: "Peck Up Your Speed"-Students analyze the relative frequency of workers'  typing speeds.

Exercise: "Rent Me"-Students calculate expected value in order to assist the Callow Corporation in making an important decision regarding leasing computer equipment.

NORMALITY

If you look at the graph of your data and notice that it looks like a bell-shaped curve, it probably  follows a normal distribution. The manufacturing industry, for example, measures volumes and weights of products. Data collected from these instances are normally distributed. In fact, if we take any process and  calculate its mean value enough times, then its mean value is also normally distributed.  This module examines the famous 'Normal Distribution' and how it can be used to provide useful information about the probability of certain events.

XII.  The Normal Distribution

       - Examples of Normally Distributed Data Variables

      - Characteristics of the Normal Distribution

      - Interpreting the  Empirical Rule

       - Components of the Normal Distribution: Probabilities and X values

       - Using the NORMDIST function in Excel to calculate probability from a normal distribution

       - Using the NORM. INV function in Excel to calculate X values related to a normal distribution

Exercise: "Fill 'R' Up"- Students help a manufacturer use the normal distribution of volumes from a cup dispenser to calculate probabilities of certain events.

ASSOCIATION AND PREDICTION

If we have data measurements for at least two variables, it is natural to ask if there is some relationship between the two. This module takes a look at the two strongest measures of association, correlation and regression, and explores how to use them to quantify the relationship.  It also examines how regression output can be used to predict future observations.

XIII. Correlation and Regression

       - Definition of Correlation and Regression

        - Relationship between Correlation and Regression

        - Correlation Coefficient: Values

       -  Examples of Correlation

       -  Interpretation of a Regression Equation

       - Step-by-Step example of How to Do a Regression Analysis

 Exercise: "Paid Sickouts"- Students use correlation and regression to help a company determine the relationship between number of sick days employees took and the wages they earned.