Business Intelligence Overview

Business Intelligence (BI) and Big Data

• BI focuses on straightforward reports and performance measures, aiding management decisions with recent trends.
• Traditional BI relies on a central server and offline data analysis, managing structured data.
• Big data operates on a distributed file system, offering real-time streaming and handling structured, semi-structured, and unstructured data.

BI vs. Data Science

• Key distinction lies in analytical techniques: Data Science emphasizes predictive analytics and data mining.
• BI utilizes standard reporting, dashboards, alerts, and queries primarily on structured data from traditional sources.
• Common questions explored in BI include historical performance, scenarios, and problem identification.

Model Planning in Data Science

• Loading data into a warehouse requires thorough understanding; local systems may emerge as departmental warehouses.
• Data is analyzed via applications, and the need for either a single model or model ensemble is established.
• Various analytic techniques are applied across industries, from linear regression in consumer goods to neural networks in telecommunications.

Common Tools for Model Planning

• SAS: High-quality interpretive models, statistical analysis, and quality plotting capabilities.
• SQL: Enables in-database analytics for predictive modeling and data manipulation.
• RapidMiner and Tableau Public: Advanced analytics with no programming required, supporting real-time data usage and sharing.

Model Building Importance

• Extracting insights informs business strategies and decisions.
• Data is divided into training (60-80%), validation (10-20%), and testing data (10-20%) to develop predictive models.

Data Warehousing Characteristics

• Central repositories for structured and semi-structured data, facilitating regular analysis.
• Key characteristics include being subject-oriented, integrated, non-volatile, and time-variant.

Types of Analytics

• Descriptive, Diagnostic, Predictive, and Prescriptive analytics offer various insights from data.
• Classification of analytics ranges from basic insights to operationalized and monetized analytics.

Evolution of Analytics

• Analytics have evolved from descriptive and diagnostic analyses (2005-2012) to predictive and prescriptive analyses (2012-present).
• The focus has shifted from historical reporting to forecasting and optimizing future actions.

Data Science vs. Data Analytics

• Data Science leans towards exploratory and predictive analytics, while Data Analytics focuses on reporting and optimization.
• Outputs differ: Data Science aims for comprehensive data products, whereas Data Analytics yields reports and dashboards.

Traditional BI vs. Big Data

• BI encompasses data mining and performance benchmarking, relying on central servers for offline analysis.
• Big data enables comprehensive analytics from multiple sources and supports large-scale data operations.

Drivers and Ecosystem of Big Data

• Emerging technologies include data devices, collectors, and aggregators, creating a new ecosystem for data usage.
• Key roles in this ecosystem demand advanced training in quantitative disciplines, with a simultaneous need for data savvy professionals.

Challenges of Big Data

• Key challenges include scale, security, schema flexibility, and ensuring continuous availability.
• Data volume increases exponentially, raising questions about usefulness and retention.
• Organizations face a talent gap, requiring skilled professionals to manage and exploit big data analytics effectively.