Business Intelligence Chapter 2

Business Intelligence, Analytics, and Data Science: A Managerial Perspective - Chapter 2

• Descriptive Analytics I: Nature of Data, Statistical Modeling, and Visualization. This chapter focuses on descriptive analytics, examining data characteristics, modelling techniques, and visualization methods used to present and analyze data.
• Nature of Data: Data is a collection of facts, often obtained from experiences, observations, or experiments. Data can be numbers, words, images, or a combination. It's the fundamental building block for information and knowledge. Data quality, including accuracy, completeness, consistency, and validity, is crucial for effective analytics. Data integrity is critical.
• Data Quality and Integrity: Data quality and integrity are essential components. Accuracy, completeness, consistency, and validity ensure reliable analysis. A strong focus is needed on data.
• Different Types of Data:
  • Structured Data: Organized, standardized format. Examples: names, dates, addresses, stock information. Follows a persistent order and easily accessed by machines.
  • Unstructured Data: Textual, imagery, voice, web content. Not easily organized by machines.
  • Semi-structured Data: Mix of structured and unstructured data. Examples: XML, HTML, JSON, log files. Partially structured.
• Metrics for Analytics Ready Data: Ensuring data quality involves assessing source reliability, content accuracy, accessibility, data security/privacy and data richness (completeness), consistency, currency (timeliness), validity (matching expected and actual data), and granularity (level of detail).
• Data Classification:
  • Categorical data: Represents types of data, divided into groups (e.g., race, gender, marital status). This can be further categorized into:
    • Nominal: Labels or names without quantitative value (e.g., colors, nationalities, types of products). Simply names.
    • Ordinal: Ordered categories where the difference between categories isn't necessarily meaningful (e.g., Likert scales, satisfaction ratings). Order matters.
  • Numerical data: Has a numerical value and is measurable. Can be categorized into:
    • Interval: Data with order and meaningful differences, but no true zero point (e.g., temperature in Celsius). Differences are meaningful.
    • Ratio: Data with order, meaningful differences, and a true zero point (e.g., height, weight, income). Zero is meaningful.
• Data Preprocessing: Essential for making data usable for analytics. Typically entails Data consolidation, Data cleaning, Data transformation, and Data reduction.
• Data Reduction Techniques:
  • Variables: Dimensional reduction and variable selection help reduce complexity by focusing on the most vital properties.
  • Cases/samples: Sampling, balancing/stratification helps ensure representation in the subset. Discretization turns continuous data into categorical intervals.
• Data Normalization: Reorganizing data to eliminate unstructured data, redundancies, and enable standardized data formats across systems.
• Statistical Modeling for Business Analytics: Overview of Statistical techniques for business.
  • Statistics: Defined as collection of mathematical theories and techniques that are used to characterize and interpret data sets.
  • Descriptive Statistics: Used to describe data. (mean, median, mode)
  • Inferential Statistics: Drawn from sample data to infer characteristics of an entire population.
• Descriptive Statistics – Measures of Central Tendency: Provides a way to summarize data.
• Arithmetic Mean: The average of a set of values.
• The Median: The middle value when the data is ordered.
• Mode: The most frequent observation.
• Descriptive Statistics – Measures of Dispersion: Describes variation in data sets.
• Range: Difference between the maximum and minimum values.
• Variance: Average of the squared deviations from the mean.
• Standard Deviation: Square root of the variance.
• Mean Absolute Deviation (MAD): Average of the absolute deviations from the mean.
• Histogram: Graph showing the frequency of data points in different ranges.
• Skewness: Measure of asymmetry in a data distribution. (negative, positive, or symmetric).
• Kurtosis: Measurement of the shape and distribution of data points. (Leptokurtic, Mesokurtic, or Platykurtic).
• Regression Modeling for Inferential Statistics: Regression techniques are tools to establish the relationships between variables. Used to forecast, predict, and further analyze data sets.
• Business Reporting: A process of collating data and making it accessible for managerial decisions. Including formats like text, tables, and charts and the manner of delivery (in-print, email, websites).
• Types of Business Reports:
  • Metric Management Reports: Helps manage business performance through key metrics.
  • Dashboard-Type Reports: Provide visual summaries of key performance indicators (KPIs) on a single page.
  • Balanced Scorecard-Type Reports: A strategic management system for internal and external outcomes, to assess financial, customer, process and learning aspects of a business.
• Data Visualization: Uses visual representations to understand, explore, and convey data, for business analysis.
• Visual Analytics: Combines information visualization with predictive analytics to create a holistic overview of data.
• Performance Dashboards: Visual displays of critical metrics for real-time monitoring and analysis. These are organized and displayed with ease and speed for effective interpretation.
• Best Practices in Dashboard Design: Designing effective dashboards involves methods for best practice, like validating the design with users, benchmarking, and selecting suitable visual elements.