Analytics Modeling and Principles Module 1

Analytics Modeling

is the process of using data, mathematics, and

statistics to help people understand situations, predict outcomes,

and make better decisions. IDENTIFICATION

Analytics Modeling

• It turns data into insights that guide actions. IDENTIFICATION

Analytics Modeling

The process involves analyzing historical data and known variables by applying mathematical or statistical techniques to identify relationships and patterns, which are then used to estimate or predict future outcomes. IDENTIFICATION

Analytics Modeling

This approach focuses on simplifying complex systems by clarifying relationships between variables, reducing uncertainty in decision-making, and transforming complexity into actionable insights. IDENTIFICATION

True

Analytics is not just about data or models, it is about making

better decisions.True or False

True

Data becomes valuable only when it helps people choose what

to do next.True or False

True

Analytics starts with a decision, not data.True or False

Define the decision problem, Translate into an analytical quesiton, analyze THE data, Generate insights, Take action, Evaluate outcomes, and refine adn repeat

What are the 7 iterative decision loop of Analytics Modeling?Enumerate

Define the decision problem

– What choice must be made?

Translate into an analytical question

What can data help

predict, compare, or explain?

Analyze the data

Use descriptive, predictive, or prescriptive

methods.

Generate insights

What does the analysis reveal?

Take action

Implement the decision.

Evaluate outcomes

Did the decision improve results?

Problem scoping

defines what problem should be

solved and how analytics can help.

Identify the right problem

is the step that ensures an analysis is meaningful and worthwhile by verifying that the results will influence decisions, that the required data exists or can be collected, and that analytics is an appropriate and suitable solution to address the problem.

symptoms

visible means

ROOT CAUSES

WHY IT HAPPENS means

Symptoms

describes outcomes

Root causes

it explains the underlying drivers

Seperating Symptoms from root causes

Analytics must go

beyond what is visible (symptoms) and investigate why it

happens (root causes).

Scope boundaries

clearly define what the analytics model will

and will not address. This prevents scope creep, manages

expectations, and aligns the project with time, data, and

resource constraints.

Setting scope boundaries

Scope boundaries clearly define what the analytics model will

and will not address. This prevents scope creep, manages

expectations, and aligns the project with time, data, and

resource constraints.

Identfying stakeholders and their needs

Different stakeholders use analytics for different decisions.

Scoping must align outputs with stakeholder needs rather than

technical complexity. Analytics is successful when decision-

makers clearly understand and can act on the results.

Translating problems into analytics task

Once the problem is scoped, it must be translated into a specific analytics

task. This step determines the type of analytics or model to be used,

ensuring the analysis directly answers the decision question.

Descriptive modeling

explains what happened or what is currently

happening.

Descriptive Modeling

It focuses on understanding historical and present data.

Mean, Median , mode

Used to summarize typical values.

Example: Average student GPA per department

Data Aggregation

Combining data to higher-level summaries.

Example: Total number of graduates per year instead of per

student.

Trend Analysis

Identifying upward or downward patterns

over time.

Example: Trend showing declining enrollment in IT programs

over 3 years.

Predictive Modeling

estimates what will happen in the future.

• It uses past data patterns to predict outcomes.

Forecasting Models

Estimate future values based on

historical data.

Example: Time-series model forecasting monthly electricity

consumption on campus.

Probability Estimation

Estimates likelihood of an event

occurring.

Example: Predicting the probability that a student will fail a

course based on attendance and quiz scores.