Types of Analytics

• There are five types of analytics: descriptive, diagnostic, predictive, prescriptive, and adaptive.

Types of Analytics

• Descriptive analytics answers "What happened?" and "What is happening?".
• Diagnostic analytics answers "Why did it happen?" and explores causes for past results, clarifying why results differ from expectations.
• Predictive analytics forecasts future events, focusing on "Will it happen in the future?" and assessing probabilities.
• Prescriptive analytics suggests actions based on predictions, answering "What should we do?" and optimizing performance within constraints.
• Adaptive analytics uses artificial intelligence for continuous learning from past and current events.

Exploratory Data Analytics

• Exploratory data analytics summarizes and explains performance and activity as initial investigation.
• It generates questions and hypotheses for further analysis and explores data.
• It involves data exploration through performance statistics, anomaly identification, and pattern detection.

Confirmatory Data Analytics

• Confirmatory analytics uses predictive and prescriptive methods, employing statistics to estimate the probability of future events.

Descriptive Analytics

• Descriptive analytics defines analytics that characterizes, summarizes, and organizes data features to improve understanding of results and underlying data.

Business Data Sources for Descriptive Analytics

• Human Resource Management Systems (HRMS) manage interactions with current and potential employees.
• HRMS data includes recruiting details, employee training, payroll, compensation, and benefits.
• Customer Relationship Management (CRM) systems manage interactions with current and potential customers.
• CRM data includes customer contact history, interests in products/services, order history, trade discounts, payment terms, and credit score.
• Supply Chain Management Systems (SCM) / Financial Reporting Systems (FRS) track the sequence of processes from production to final delivery.
• SCM data includes active vendors, order details, and demand schedules for product components.

Financial Reporting Systems

• Financial Reporting Systems (FRS), capture and measure financial transactions and communicate financial performance.
• Financial Statements are reports on financial results, condition, health, and cash flows.

Descriptive Analytics Techniques

• Descriptive analytics techniques address "What happened?" by characterizing, summarizing, and organizing data.
• These include measures like mean, median, mode, minimums, maximums, standard deviation, quartiles/deciles and counts to explain the data.
• Totals, sums, subtotals, bar charts, ratio, vertical, and horizontal analyses are also components of descriptive analysis.
• Percentage change, pivot tables, and histograms are also key techniques.
• Descriptive analytics is used for identifying patterns and relationships in historical data.

Horizontal Analysis

• Calculates and displays the dollar and percentage change in financial statement line items from one period to the next.

Vertical Analysis

• Examines financial performance over time by relating financial information to a relevant figure or base.

Diagnostic Analysis

• Diagnostic analytics investigates reasons for past results that descriptive data does not answer.
• It identifies anomalies, outliers, previously unknown linkages, patterns, relationships between variables, performs drill-down detailed analytics, and statistical analyses.
• An anomaly is data that deviates from the expected norm, and an outlier is an observation outside of its expected distribution.

Deterministic vs. Probabilistic Analyses

• Deterministic analysis considers the impact of a single risk scenario.
• Probabilistic analysis considers all possible scenarios, their likelihood, and associated impacts, recognizing an element of chance.

Confirmatory Analytics

• Confirmatory analytics are predictive and prescriptive, using techniques and statistics to evaluate the likelihood or probability of future events.

Predictive Analytics

• Predictive analytics determines whether something will happen in the future by identifying patterns in historical data.

Predictive Analytics Techniques

• Classification separates a sample into groups or classes.
• Regression predicts a specific outcome based on independent variables.
• Time series analysis predicts future values based on past values.
• Classifying groups can be done by identifying and marketing to customer segments, extending or denying loans, and identifying fraud.
• Customer segmentation can be categorized by communication, human, physical, and sensory touchpoints.
• Segmentation can be based on sales levels and types of purchases.
• Banks evaluate creditworthiness to decide whether to extend loans, using credit scores, work history, debt levels, and home ownership as factors.

Predictive Analytics Technique: Regression

• Regression allows analysts to predict a dependent variable outcome based on independent variable inputs.
• Assess if independent variables are related to dependent ones.

Interest Rates and Social Media Sentiment Examples

• Determining the interest rates given to a borrower.
• The dependent variable is the interest rate given.
• Independent variables are borrower income, debt level, credit history, and housing status.

Time Series

• This analysis predicts future values of a variable based on its past values.
• Accountants measure the past to forecast the future.

Base Rates & Base Rate Fallacy

• Base rates are event probabilities based on historical averages.
• The base rate fallacy occurs when predictions underweight base rates, relying on new information instead.

Prescriptive Analytics

• Identifies optimal options given constraints, making specific recommendations.

Prescriptive Analytics Examples

• What-If Analysis
• Cash Flow Analysis
• Scenario Analysis
• Goal-Seek Analysis
• Sensitivity Analysis

Prescriptive Analytics Usefulness

• Helpful in identifying options given constraints and changing conditions.
• Constraints include limited capital resources.
• Changing consumer preferences, tax laws, interest rates, and demand are factors.

What if (MARGINAL) Analysis

• Evaluates how potential changes in inputs may affect output and analyzes the cost or benefit of specific decisions.
• Marginal analysis refers to the cost or benefit of the next unit.
• This analysis can apply financial modeling, representing outcomes to a decision or future event.
• Manufacturer decisions involve expanding product lines or increasing volume.

Cash Flow Analysis

• A prescriptive analytic technique, evaluates cash flows for potential expenditures.

What if Scenario Analysis

• Analyzes future events by considering potential outcomes
• It presents a range of alternative incomes
• It computes the expected overall impact on the company based on the joint probability of multiple events

Goal Seek Analysis

• Tells us what input will be needed to reach a desired outcome
• Helps calculate backwards to understand the input needed to achieve a certain outcome

Sensitivity Analysis

• Evaluates outcomes based on uncertainty regarding the inputs
• If fuel prices increase by 10%, 15% or 20%, what will be the impact on net income for FedEx?

Prescriptive Analytics Techniques

• There are several different techniques one can use: Marginal/Incremental analysis, cash-flow analysis, scenario analysis, goal-seek analysis, sensitivity analysis