Introduction to Data Analytics for Accounting

Accounting is shifting from sample-based checking to full-population analysis, and this book is the standard on-ramp. It suits accounting students, auditors, tax and advisory professionals, and managers who want to read data critically. You get frameworks you can apply immediately, plus honest warnings about dirty data, bias, and professional skepticism.

1. 1Chapter 1

   The IMPACT cycle: six steps for any analytics problem

   Data analytics means evaluating data to draw conclusions that address business questions, so the question comes before the data. The book hands the reader a sequence instead of a starting point in the spreadsheets: Identify the questions. Master the data. Perform the test plan. Address and refine results. Communicate insights. Track outcomes. The pile of tabs becomes a decision because the steps run in order.

   The idea

   The cycle's power is its starting point. Most failed analytics projects begin with data and hunt for a purpose; this one begins with a business question and works backward. That ordering keeps technical work tied to decisions, which is why the book treats it as the standard approach for every problem, from audit tests to budgeting. Skipping a step, especially tracking outcomes, quietly breaks the loop.

   Try this: Before opening any dataset this week, write the business question in one sentence, then map it to the six IMPACT steps.

2. 2Chapter 1

   Four questions, four kinds of analytics

   Why did margins move? Diagnostic analytics digs in. Will it happen next quarter? Predictive analytics estimates it. What should we do? Prescriptive analytics recommends. Naming the type before you start prevents scope creep and sets stakeholder expectations.

   The idea

   Naming the type before you start prevents scope creep and sets stakeholder expectations.

   Try this: Label your next three analytics requests as descriptive, diagnostic, predictive, or prescriptive before estimating effort.

3. 3Chapter 1, Chapter 11

   Enron: the fraud data could have caught

   The book asks what a data-equipped analyst could have done: pull footnote disclosures, compute Enron's debt-to-equity ratio, and compare it against industry peers. Better data access, the book argues, could have flagged the fraud earlier.

   The idea

   Analytics changes the unit of attention from a single filing to a comparison across companies and time.

   Try this: Pick one public company and compare its debt-to-equity ratio to three industry peers using their XBRL filings.

4. 4Chapter 1, Chapter 19

   The Big 4 bet big on analytics

   Deloitte, EY, KPMG, and PwC have all made significant investments in data analytics capabilities and tools. Deloitte built Omnia, an AI audit tool. PwC's Halo tests 100% of transactions.

   The idea

   The firms already run on analytics, and fluency matters for an audit career.

   Try this: Look up one audit tool from a Big 4 firm and note which audit step it automates.

5. 5Chapter 1, Chapter 6, Chapter 17

   Sarah tests 50,000 invoices, not 50

   Sarah tests Cloud 9's 50,000 sales invoices in full with data analytics and finds duplicate invoices. Full-population testing replaces sampling in some cases, and anomalies surface in days rather than weeks.

   The idea

   The auditor's role shifts from designing samples to interpreting exceptions.

   Try this: Run a duplicate check on one month of your expense or invoice data this week.

6. 6Chapter 2

   80% of analytics is cleaning dirty data

   The book notes analysts often spend up to 80% of their time preparing and cleaning data. Data comes structured, like rows and columns in a database, or unstructured, like emails and social media posts, with semi-structured in between.

   The idea

   Budgeting time for cleansing matters, and the book puts this chapter early.

   Try this: Audit one dataset you own for the five dirty-data problems and fix the duplicates first.

7. 7Chapter 2, Chapter 3

   Four Vs and the data warehouse

   Big data gets measured on four Vs: Volume, Velocity, Variety, and Veracity. Data splits by source into internal versus external, and by nature into financial versus non-financial. To make it usable, companies build a data warehouse, a central repository integrating data from multiple sources. Data lakes go further, holding both structured and unstructured data, while warehouses hold structured data built for reporting.

   The idea

   The Vs and the warehouse-versus-lake distinction shape how a question gets answered.

   Try this: Classify your company's last analytics request by the four Vs and note which V limited the answer.

8. 8Chapter 4

   50,000 alumni, a $250 mean, and a misleading average

   The book examines a university alumni donation database with 50,000 alumni, a mean donation of $250, a median of $100, and a standard deviation of $400, where the top 5% of donors contribute 60% of total donations.

   The idea

   Summary statistics need care when distributions are skewed.

   Try this: Compute mean, median, and standard deviation for one of your datasets and ask which figure a headline should quote.

9. 9Chapter 4, Chapter 6

   Tufte's rule: maximum ideas, minimum ink

   The book presents Edward Tufte's visualization principles and his data-ink ratio concept.

   The idea

   Data-ink ratio shapes how much a chart communicates.

   Try this: Take one of your charts and delete every element that is not data, then see if the message survives.

10. 10Chapter 4, Chapter 6

    Ads, sales, and an R-squared of 0.78

    The book pairs a regression of monthly sales on advertising spend with the ice cream example, where ice cream sales and drowning deaths rise together in summer yet one does not cause the other. Correlation is not causation.

    The idea

    Correlation is not causation, and hidden variables can drive both series.

    Try this: For one relationship you plan to act on, list two alternative explanations before trusting the coefficient.

11. 11Chapter 5, Chapter 3, Appendix A

    Excel does more than you think

    The book shows PivotTables summarizing sales by product category and region, VLOOKUP merging data from two tables, conditional formatting highlighting what matters, and appendices walking through sorting, filtering, and PivotTable creation step by step.

    The idea

    Excel can carry a surprising share of analytics work, and its operations transfer to heavier tools.

    Try this: Rebuild one summary report this week as a PivotTable instead of manual formulas.

12. 12Chapter 7, Chapter 6, Chapter 8

    Benford's Law catches the invoice that lied

    The book describes an invoice dataset where the leading digit 1 showed up only 12% of the time, a red flag for manipulation, against the roughly 30.1% Benford's Law expects in naturally occurring datasets.

    The idea

    Comparing expected digit frequencies against actual ones tells you where to look next.

    Try this: Extract leading digits from one invoice or expense list and compare the digit 1 frequency to 30.1%.

13. 13Chapter 7, Chapter 6, Chapter 8

    The AICPA blesses audit analytics

    The AICPA's Guide to Audit Data Analytics, published in 2017, gives auditors a playbook for implementation, and the Audit Data Standards standardize data fields so client data arrives audit-ready.

    The idea

    Standards and standardized data fields shape how analytics enters the audit.

    Try this: Read one section of the AICPA's audit data analytics guidance and map it to your next audit step.

14. 14Chapter 7, Chapter 10

    Lisa finds the faulty machine

    Lisa, a production manager at Broadway Industries, digs into material usage data and finds a 15% unfavorable variance caused by a faulty machine, which gets repaired. In a separate example, a labor variance of $15,000, budgeted $100,000 against actual $115,000, breaks into an $8,000 price variance and a $7,000 quantity variance.

    The idea

    Splitting variances into components points toward causes.

    Try this: Take one variance in your budget this month and split it into price and quantity components.

15. 15Chapter 8, Chapter 7

    Balanced scorecard meets analytics

    The balanced scorecard becomes sharper when fed with data, with financial and non-financial measures refreshing together.

    The idea

    Feeding the scorecard with data changes how its measures behave.

    Try this: Pick one balanced scorecard measure and set up a weekly automated refresh for it.

16. 16Chapter 9, Chapter 11

    Ratios, aging, and XBRL

    The SEC requires all public companies to file financial statements in XBRL format, and those tags allow automated comparison of ratios across companies.

    The idea

    Standardized tags make peer comparison a query.

    Try this: Pull two competitors' XBRL filings and compare one ratio automatically using a free SEC viewer.

17. 17Chapter 10, Chapter 12

    The IRS scores your return

    The IRS uses the Discriminant Function System, assigning DIF scores to returns, with a score above 100 indicating a high probability of audit.

    The idea

    Audit selection is statistical, which changes how returns get prepared.

    Try this: Before filing, run one analytics check on your return for common compliance errors like mismatched forms.

18. 18Chapter 11, Chapter 10

    Maria predicts which clients will breach covenants

    Maria, a partner, uses predictive analytics to identify clients at high risk of missing loan covenants, and the firm reaches out proactively with advisory services.

    The idea

    Prediction can shift an engagement toward forward-looking advice.

    Try this: Identify one metric that predicts trouble for a client or business unit and set a threshold alert for it.

19. 19Chapter 11, Chapter 14

    Storytelling with data: context, conflict, resolution

    The book introduces a storytelling framework with three elements: context, conflict, and resolution, and recommends starting with the 'so what' question.

    The idea

    Framing results as a story helps non-technical stakeholders follow them.

    Try this: Rewrite your next findings slide to open with the 'so what' in one sentence before any chart.

20. 20Throughout, Appendix

    Skepticism, ethics, and the tools you'll practice on

    The book urges professional skepticism with analytics results and attention to data privacy and algorithmic bias, alongside hands-on exercises in Excel, Tableau, and Power BI, including a Tableau tutorial building a bar chart of sales by region filtered to sales above $10,000 and a SQL tutorial querying Customers and Orders tables.

    The idea

    Skepticism and practice advance together with the tools.

    Try this: For one automated report you trust, list two ways its underlying data could be biased or incomplete.

Accounting students preparing for a profession that now expects analytics fluency, and practicing auditors, tax specialists, and finance managers who want structured frameworks and tool practice. Also useful for anyone advising clients with data.

FAQ