Data Analytics in Accounting

The Four V’s of Big Data

  • Big data: Massive amounts of data companies capture, store, and analyze.
    • Data volume: Amount of data created/stored.
    • Data velocity: Speed at which data is created/stored.
    • Data variety: Different forms data can take.
    • Data veracity: Quality/trustworthiness of data.

Analytics Mindset

  • Analytics mindset: Correct use of data and analysis for decision making.
    • Includes:
    • Asking the right questions.
    • Extracting, transforming, and loading relevant data.
    • Applying appropriate data analytic techniques.
    • Interpreting and sharing results.

Ask the Right Questions

  • Good questions establish "SMART" objectives:
    • Specific: direct and focused.
    • Measurable: amenable to data analysis.
    • Achievable: answerable, leading to action.
    • Relevant: relates to objectives.
    • Timely: defined time horizon.

Extract, Transform, Load (ETL) Process

  • ETL: Extracting, transforming, and loading data.
    • Time-consuming.
    • Can be automated.

Extracting Data

  • Steps:
    • Understand data needs and available data.
    • Perform data extraction.
    • Verify quality and document.
  • Data organization:
    • Structured: Highly organized (e.g., accounting data).
    • Semi-structured: Not structured enough for databases (e.g., CSV).
    • Unstructured: Most public data (e.g., images, tweets).

Transforming Data

  • Steps:
    • Understand data and desired outcome.
    • Standardize, structure, and clean.
    • Validate quality and verify requirements.
    • Document process.

Loading Data

  • Considerations:
    • Format acceptable to receiving software.
    • Understand how the new program interprets data.
    • Update or create a data dictionary.

Data Analytic Techniques

  • Four categories:
    • Descriptive: Understand the past; "what happened?"
    • Diagnostic: Why it happened; "why did this happen?"
    • Predictive: Predict the future; "what might happen?"
    • Prescriptive: Recommend action; "what should be done?"

Interpreting Results

  • Potential issues:
    • Confusing correlation with causation.
    • Systematic biases in interpretation.

Sharing Results

  • Data storytelling: Translating complex analyses for better decision-making.
    • Remember objectives.
    • Consider audience.
    • Use effective visualizations.

Data Visualization

  • Graphical representation to convey meaning.
    • Principles:
    • Right visualization type.
    • Simplified presentation.
    • Emphasize important aspects.
    • Ethical representation.

Automation

  • Machines perform tasks automatically.
  • Robotic Process Automation (RPA): Automates tasks across applications.
    • Used to automate ETL tasks.

Data Analytics Limitations

  • Not always the right tool.
    • Reliable data may not exist.
    • Human judgment may be needed.
    • Importance of intuition, expertise and ethics.