Week 6

WEEK 6: DATA AND APPLIED INTELLIGENCE

Introduction

  • Subject: BTE 210: Fundamentals of Business Technology & Innovation

  • Instructor: Dr. Alex Korogodsky (alex.korogodsky@miami.edu)

  • Focus: Developing innovative ideas and principled leaders that transform global business and society, particularly at the crossroads of the Americas.

PART 1: BUSINESS INTELLIGENCE

The Need for Data Analysis

  • Critical for Managers:

    • Managers need to track daily transactions to evaluate business performance.

    • Development of strategies must align with organizational goals through operational databases.

    • Data analysis offers insights into short-term tactical evaluations and strategies.

  • Business Intelligence (BI):

    • A comprehensive collection of tools and processes that includes:

    • Capture, collect, integrate, store, and analyze data.

    • Generate actionable information to support decision-making.

    • Transformation Process:

    • Data into information: Through organization and analysis.

    • Information into knowledge: Applying context and interpretation.

    • Knowledge into wisdom: Facilitating informed decision-making.

Managerial Decision-Making

  • Roles of Managers:

    1. Interpersonal Roles: Interacting with others to fulfill organizational objectives.

    2. Informational Roles: Gathering and disseminating information.

    3. Decisional Roles: Making choices that affect the organization.

  • Types of Decision Structures:

    • Problem Structure:

    • Highly Structured: Examples include order entry.

    • Semi-Structured: Examples include loan approval.

    • Highly Unstructured: Example includes determining a new store location.

The Information Gap

  • Definition: The shortfall between gathering information and utilizing it effectively for decision-making.

  • Key Issues:

    • Many firms have inadequate data warehouses.

    • Business Analysts often spend 80% of their time gathering, cleaning, and preparing data, leaving only 20% for analysis and real decision-making.

  • Proposed Paradigm Shift:

    • Reverse the time allocation to 20% preparation and 80% analysis.

  • Goal of Business Intelligence (BI):

    • To bridge the information gap by optimizing data analysis processes.

How BI Works

Data Models and Systems

  • Types of Systems Utilized:

    • ERP (Enterprise Resource Planning): Integrates core business processes.

    • CRM (Customer Relationship Management): Manages company interactions with current and potential customers.

    • SCM (Supply Chain Management): Manages the flow of goods and services.

    • Data Warehousing: Central repository for stored data.

  • Methods of Data Utilization:

    • Getting Data In: Through various enterprise and departmental systems.

    • Getting Data Out: Analyzing and visualizing information for insights.

BI Applications

  • Types of Analyses:

    • Multidimensional Analysis (OLAP): Allows for "slicing and dicing" of data stored in a dimensional format for deeper insights.

    • Data Mining: Seeks to discover patterns in large datasets using statistical methods and artificial intelligence.

    • Generates insights from large databases and uncovers previously unknown features.

    • Decision Support Systems: Combine data models to address semi-structured and unstructured problems.

    • Examples include:

      • Customer Analytics

      • Human Capital Productivity Analysis

      • Supply Chain Analytics

      • Behavior Analytics

BI Implementation Steps

  1. Collect Operational Data: Start by gathering the necessary operational data.

  2. Aggregate Data: Transform operational data (What happened?) into decision support data (Why it happened?).

  3. Analyze Data: Generate information from decision support data to make sense of it.

  4. Present Information: Show the analyzed information to end-users to assist in decision-making.

  5. Make Business Decisions: The decisions lead back into the system, generating additional data.

  6. Monitor Results: Evaluate the outcomes of business decisions through ongoing data collection.

Data for Decision Making

  • Quality and Source of Data:

    • Raw data must be collected either manually or via instruments, emphasizing the critical nature of quality.

    • Quality issues often exposed when data is summarized.

  • Common Problems Encountered:

    • Data may be neglected or poorly handled, impacting its usefulness.

Visualization

  • Dashboards:

    • Offer rapid access to timely information and management reports in a user-friendly format.

PART 2: APPLIED INTELLIGENCE

Understanding Trends vs. Historical Data

  • Trends: Understanding future potential and operational adjustments.

    • Example questions:

    • "How to increase growth rate by 10%?"

    • "Should I invest $100 to retain employees vs. $150 for termination?"

  • Focus of Market Leaders: Predicting rather than merely reporting what has previously occurred.

The Evolution of Enterprise Intelligence

  • Value Complexity Matrix:

    • High business value vs. Low complexity is ideal for decision-making processes.

    • Stages of Evolution:

    • Reporting (What happened?)

    • Analysis (Why did it happen?)

    • Monitoring (What is happening now?)

    • Prediction (What will happen?)

    • Optimization (What is the best outcome?)

  • Enterprise Intelligence Stack Overview:

    • Includes:

    • Artificial Intelligence & Cognitive Processes

    • Prescriptive Analytics and Decision Optimization

    • Machine Learning and Predictive Modeling

    • Intelligent Automation and Robotic Process Automation (RPA)

    • Business Intelligence with multidimensional analysis and visualization

    • Reporting through queries and ad-hoc reporting.

Intelligent Automation

  • Applications and Benefits:

    • Optimizes operations across functions like Finance and Marketing via human-centered automation strategies.

    • Engages customers and empowers employees using digital assistants, potentially reducing operational costs by 30%-60%.

    • Automates back-office processes in Finance, HR, IT, Procurement, and Sales.

  • Examples of Intelligent Automation Solutions:

    • Transferring data across systems (e.g., updating customer files).

    • Intelligent call center agents utilizing natural language processing (NLP).

Machine Learning

Definition and Algorithms

  • Machine Learning: Systems that learn from data, utilizing algorithms to build a mathematical model from sample data (training data).

  • Capable of making predictions or decisions independently of explicit programming.

Applications of Machine Learning

  • Real-World Applications Include:

    • Optical Character Recognition: Recognizes printed and handwritten text using previous examples.

    • Face Recognition: Identifies faces in images through neural networks.

    • Sentiment Analysis: Studies emotional states and subjective information from text.

    • Fraud Detection: Pinpoints potentially fraudulent credit card transactions.

Predictive Modeling

Overview

  • Applications Across Industries:

    • General predictions on employee performance or supply chain efficiency.

    • Specific applications in various sectors such as energy, manufacturing, retail, and finance.

Predictive Modeling Steps:

  1. Prepare Data: Ensure data is formatted and relevant.

  2. Analyze Data: Evaluate data for completeness and quality.

  3. Build Models: Create predictive models.

  4. Assess Performance: Employ a confusion matrix to evaluate model predictions.

    • Example Confusion Matrix:
      | Actual/Predicted | Top Performer | Low Performer |
      |------------------|---------------|---------------|
      | Top Performer | 270 | 120 |
      | Low Performer | 90 | 550 |

    • Accuracy Results: Top performer - 69.23%, Low performer - 85.93%, Overall model accuracy - 79.61%.

Example: Intelligent Recruiting with AI

  • Functionality:

    • Matches candidate skill profiles to job descriptions for improved recruitment efficiency, allowing clear insights into candidate potential.

  • Data Enrichment:

    • Ability to incorporate external data for a fuller understanding of candidate profiles.

No-Code AI Model Building

Overview of Microsoft Lobe AI

  • *Functionality Profile:

    • Automatically selects appropriate AI architecture for user projects, simplifying image classification, object detection, and data classification tasks.

  • Steps for Building Models:

    1. Data collection and labeling through images.

    2. Training models with live evaluation for strengths and weaknesses.

    3. Model deployment and continuous improvement based on feedback.

Machine Learning Career Overview

Salary Insights:

  • Entry Level (0-4 years): Average salary is approximately $97,090, potentially rising to $130,000 with bonuses.

  • Mid-Level (5-9 years): Average salary is $112,095, with potential to reach $160,000.

  • Senior Level (10+ years): Average salary is $132,500, and can exceed $181,000 annually with bonuses.