Analytics- Exam 1

business analytics

The practice and art of bringing quantitative data to bear on decision-making.

• need to achieve a competitive advantage

  • unique, hard to copy, valuable to the customer

• A field of study that uses data, computers, statistics and mathematics to solve business problems

• Helps managers gain improved insight about their business operations and make better, fact-based decisions.

The Four V’s of Big Data

1. volume

Amount or quantity of data (how much)

• An example of a high-____ data set would be all credit card transactions on a day within Europe

The Four V’s of Big Data

2. velocity

How fast the data will be generated (time)

• An example of a data that is generated with high ____ would be Twitter messages or Facebook posts.

The Four V’s of Big Data

3. variety

Different types of data

• An example of high ____ data set would be the CCTV audio and video files that are generated at various locations in a city

The Four V’s of Big Data

4. veracity

Accuracy and trustworthiness of the data

• An example of a high _____ data set would be data from a medical experiment or trial

Descriptive analytics: “ What happened?”

The use of data to understand past and current business performance and make informed decisions.

• Do not tell managers what to do

• Data visualization, Pivot Tables, Descriptive statistics

Predictive analytics: “What may happen?”

Relationship between variables and predict the future by examining historical data, detecting patterns or relationships in these data, and then extrapolating these relationships forward in time.

Prescriptive analytics: “What should happen”

Identify the best alternatives to minimize or maximize some objective

For a supervised task (classification or prediction), randomly partition the dataset into three parts:

training, validation, and test datasets.

Data Partitions for supervised methods

• Training Partition

It is typically the largest partition and contains the data used to build (create) the various models we are examining. The same training partition is generally used to develop multiple models.

Data Partitions for supervised methods

• Validation Partition

It is used to assess the predictive performance of each model so that you can compare models and choose the best one.

Data Partitions for supervised methods

• Test Partition

It is used to assess the performance of the chosen model with new data.

Prescriptive Analytics

• Optimization

  • objective function

  • constraints

  • optimal solution

Finding values of decision variables that minimize (or maximize) something such as cost (or profit).

• The equation that minimizes (or maximizes) the quantity of interest.

• Limitations or restrictions.

• Values of the decision variables at the minimum (or maximum) point

Examples of Applications

• pricing

Setting prices for consumer and industrial goods, government contracts, and maintenance contracts

Examples of Applications

• customer segmentation

Identifying and targeting key customer groups in retail, insurance, and credit card industries

Examples of Applications

• merchandising

Determining brands to buy, quantities, and allocations

Examples of Applications

• location

Finding the best location for bank branches and ATMs, or where to service industrial equipment

Examples of Applications

• social media

Understand trends and customer perceptions; assist marketing managers and product designers

Example: A Prescriptive Pricing Model

• A firm wishes to determine the best pricing for one of its products in order to maximize profit.

• Analysts determined the following predictive model:

  • Demand = -2.9485(price) + 3240.9

  • Total revenue = (price)(Demand)

  • Cost = 10(Demand) + 5000

  • Profit= revenue - cost

• Decision Variable => price (P) 

• Objective function => maximize the profit 

• Constraints => competitor prices and production capacity demand 

• Identify the price that maximizes profit, subject to any constraints that might exist

Predictive analytics gives the demand equation; prescriptive analytics uses that equation to find the best price.

Excel:

demand= 3240.9 - 2.9485×500 (estimated price)= 1766.65

revenue= 500(1766.65)= 883,325

cost= =10×1766.65+5000= 22,660.50

profit= 883,325 - 22,660.50= 860,658.50

use solver: max, objective: profit, changing cell $500

• logical constraints: price >=0 and demand >=0

• = $554.58 and max profit now = 869,443

3 criteria for setting up a prescriptive analytics / optimization model

1. Understand the case — figure out the business problem and what the company is trying to accomplish, such as maximizing profit or minimizing cost.

2. Find the values/decision variables — identify what you are trying to determine. For example:

   • How many units should we produce?

   • What price should we charge?

   • How much should we spend on advertising?

3. Identify logical constraints — rules that must make sense mathematically or in real life. Examples:

   • Units produced must be ≥ 0

   • You cannot produce half of a product if whole units are required

   • Number of employees cannot be negative

Goals for Spreadsheet Design

• communication

• reliability

• auditability

• modifiability

• A spreadsheet's primary business purpose is communicating information to managers.

• The output a spreadsheet generates should be correct and consistent.

• A manager should be able to retrace the steps followed to generate the different outputs from the model in order to understand and verify results. (most important)

• A well-designed spreadsheet should be easy to change or enhance in order to meet dynamic user requirements.

Spreadsheet Design Guidelines

• Organize the data, then build the model around the data.

• Do not embed numeric constants in formulas.

• Things which are logically related should be physically related.

• Use formulas that can be copied

• Column/rows totals should be close to the columns/rows being totaled.

• The English-reading eye scans left to right, top to bottom.

• Use color, shading, borders and protection to distinguish changeable parameters from other model elements.

• Use text boxes and cell notes to document various elements of the model.

Descriptive analytics

examine historical data for similar products (prices, units sold, advertising)

Predictive Analytics

predict sales based on price

Prediction vs. Classification

different based on variable of interest

Prescriptive analytics

find the best sets of pricing and advertising to maximize sales revenue

Predictive analytics

• supervised vs unsupervised

• Supervised (We must have data available in which the value of the outcome of interest is known. These data are called Labeled Data)

  • prediction and classification

  • These methods “learn” or are “trained” about the relationship between predictor variables and the outcome variable

• Unsupervised (If the available data has no outcome variable to predict or classify. These data are called Unlabeled Data)

  • “what goes together” and segmentation

  • These is “no learning” from cases where an outcome variable is known

Predictive analytics

• Prediction

We are trying to predict the value of a numerical value

Example:

• predict house price using size, number of rooms

• predict sales using price, coupons and advertising

Predictive analytics

• classification

We are trying to predict a class (e.g. purchaser or no purchaser)

Example:

• the recipient of an offer can respond or not respond

• an applicant for a loan can repay on time, late, or declare bankruptcy

• a credit card transaction can be normal or fraudulent

Predictive analytics

• “What goes with what”

Association rules are designed to find association patterns between items in large databases

Example:

• there is an association between beer and diaper

• medical researcher wants to learn what symptoms appear together

• word combinations that appear too often might indicate plagiarism

Predictive analytics

• segmentation

Is used to segment the data into a set of HOMOGENEOUS clusters of records for the purpose of generating insight

Example:

• market segmentation

• finding groups of similar firms based on measures such as growth rate, profitability, market size, product range

Benefits and Challenges

Benefits

• reduced costs

• better risk management

• faster decisions

• better productivity

• remain competitive

• enhanced bottom-line performance

Challenges

• lack of understanding how to use analytics

• insufficient analytical skills

• difficulty in getting good data and sharing information

Steps in implementing an LP model in a spreadsheet

1.Organize the data for the model on the spreadsheet.

2.Reserve separate cells in the spreadsheet for each decision variable in the model.

3.Create a formula in a cell in the spreadsheet that corresponds to the objective function.

4.For each constraint, create a formula in a separate cell in the spreadsheet that corresponds to the left-hand side (LHS) of the constraint.

How solver views the model

• objective cell

• variable cells

• constraint cells

• the cell in the spreadsheet that represents the objective function

• the cells in the spreadsheet representing the decision variables

• the cells in the spreadsheet representing the LHS formulas on the constraints

5 Steps in Formulating LP Models

1. Understand the problem.

2. Identify the decision variables.

   • X₁=number of Aqua-Spas to produce

   • X₂=number of Hydro-Luxes to produce

3. State the objective function as a linear combination of the decision variables.

• MAX: 350X₁ + 300X₂

4. State the constraints as linear combinations of the decision variables

5. Identify any upper or lower bounds on the decision variables.

• X₁ >= 0

• X₂ >= 0

DV always start as

empty cells

ex. Aqua spa: number of units to be produced 

LHS vs RHS of constraint

LHS = Left-Hand Side → usually the actual amount used, produced, earned, etc. It is typically calculated with an Excel formula.

RHS = Right-Hand Side → usually the limit, requirement, or available amount given in the problem.

For example, if a company has 1,566 labor hours available:

9X1​+6X2​≤1566

When to max. vs. min. for a problem

• Maximize when you want the biggest possible value, such as profit, revenue, production, sales, or return.

• Minimize when you want the smallest possible value, such as cost, time, distance, waste, labor hours, or risk.

For example, if a problem says:

• “Determine how many hot tubs to produce to earn the greatest total profit.”

  • You would set Solver to Max because the objective is profit.

If it says:

• “Find the shipping plan that results in the lowest total transportation cost.”

  • You would set Solver to Min because the objective is cost.

Supervised

• prediction and classification example

Unsupervised

• “what goes together” and segmentation example

Supervised

prediction example: linear regression

classification example: logistic regression, K-nearest neighbors

Unsupervised

“what goes together” example: association rules

segmentation example: cluster analysis