Big Data and Analytics Quiz

Big Data

• Data volume: The amount of data collected and stored by an organization.
• Data velocity: The speed at which data is created and stored.
• Data variety: The different forms of data captured.
• Data veracity: The quality and trustworthiness of data.

Analytics Mindset

• An analytics mindset uses data and analysis for decision making.
• EY defines analytics mindset as:
  • Asking the right questions.
  • Extracting, transforming, and loading relevant data.
  • Applying appropriate data analytic techniques.
  • Interpreting and sharing results with stakeholders.

Ask the Right Questions

• A good data analytic question is:
  • Specific: Focused and direct for a meaningful answer.
  • Measurable: Amenable to data analysis with measurable input.
  • Achievable: Answerable and actionable.
  • Relevant: Related to organizational objectives or the situation.
  • Timely: With a defined answer timeframe.

Extract, Transform, and Load (ETL) Relevant Data

• ETL is the process of extracting, transforming, and loading data.
• Often the most time-consuming part of the analytics mindset process.
• Repetitive ETL processes can be fully automated, appearing as a unified step.

Extracting Data

• Three steps:
  • Understand data needs and availability.
  • Perform the data extraction.
  • Verify extraction quality and document the process.

Enterprise Data Warehouse Components

• Shows the components of a Data Warehouse.

Transforming Data

• Four steps:
  • Understand the data and desired outcome.
  • Standardize, structure, and clean the data.
  • Validate data quality and verify it meets requirements.
  • Document the transformation process.

Loading Data

• Important considerations:
  • The transformed data must be in a format acceptable to the receiving software.
  • Understand how the new program interprets data formats.
  • Update or create a new data dictionary after successful loading.

Apply Appropriate Data Analytic Techniques

• Four categories:
  • Descriptive analytics: Information about the past ("what happened?").
  • Diagnostic analytics: Builds on descriptive analytics to answer "why did this happen?"
  • Predictive analytics: Information about the future ("what might happen?").
  • Prescriptive analytics: Information with recommendations of what should be done ("what should be done?").

Ernst & Young Foundation Recommended Data Analytics Skills

• Lists recommended skills for data analytics.

Interpreting Results

• Potential misinterpretations:
  • Correlation vs. Causation: Correlation means two things happen together, causation means one thing causes the other..
  • Systematic biases in interpreting results, especially in Psychology research.

Sharing Results

• Data storytelling: Translating complex analytics into understandable terms for better decisions.
• To tell a successful data story, you need to:
  • Remember the question that initiated the analytics process.
  • Consider the audience.
  • Use data visualizations.

Data Visualization

• The use of graphical representation to convey meaning.
• Good principles of visualization design include:
  • Choosing the right type of visualization.
  • Simplifying data presentation.
  • Emphasizing important information.
  • Representing data ethically.

Automation

• Applying machines to automate tasks previously done by humans.
• Robotic process automation (RPA): Software that can be programmed to automate tasks across applications like a human.
• Companies use RPA to automate tasks within analytics processes.
• RPA can automate ETL tasks.

Data Analytics is Not Always the Right Tool

• Data analytics is not always the best tool to reach the best outcome.
• Reliable data may not exist for all questions.
• Human judgment can account for sentiment factors that are difficult to measure.
• Data helps with better decisions, but intuition, expertise, ethics, and other knowledge sources not easily quantifiable are important for performance.

Real-Time Analytics/Decision Requirements

• Real-time analytics can be applied to:
  • Product recommendations: Provide relevant and compelling recommendations.
  • Understanding customer behavior: Learn why customers switch to competitors and how to counter their offers.
  • Friend invitations: Improve marketing effectiveness of a promotion while it's still in play.
  • Preventing Fraud: Prevent fraud as it is occurring and more proactively.

Key Terms

• Big data
• Data volume
• Data velocity
• Data variety
• Data veracity
• Dark data
• Data swamps
• Metadata
• Data owner
• Flat file
• Delimiter
• Text qualifier
• Mindset
• Analytics mindset
• ETL process
• Structured data
• Unstructured data
• Semi-structured data
• Data marts
• Data lake
• Data dashboard
• Data storytelling
• Data visualization
• RPA
• Automation
• Bot