A3 Subset: Audit Data Analytics Categorization

Question 1

What type of analytic is clustering? (Descriptive, diagnostic, predictive, or prescriptive)?

Answer 1

Diagnostic (“why did it happen?”) — Clustering groups data points with similar characteristics together. Auditors use it to uncover patterns or anomalies that explain why certain transactions or accounts behave differently from others. It asks “Why are these data clustered?”

Question 2

What type of analytic is data sorting and filtering? (Descriptive, diagnostic, predictive, or prescriptive)?

Answer 2

Descriptive (“what happened?”) — Sorting and filtering organizes or narrows data so you can see what is in it. It describes the dataset without trying to explain causes or predict outcomes.

Question 3

What type of analytic is binning? (Descriptive, diagnostic, predictive, or prescriptive)?

Answer 3

Descriptive (“what happened?”) — Binning groups continuous data into ranges or buckets (e.g., ages 0–10, 11–20). It’s used to summarize and understand the distribution of data.

NOTE: Binning is descriptive, because you’re simply “binning” what’s already there (like manually sorting emails into the correct folder). This is different from classification, which is predictive - classification categorizes future items based on current characteristics (like marking emails as spam, so that all future emails will be sorted the same way - you’re predicting that any future email from that address will also be spam).

Question 4

What type of analytic is natural language processing (NLP)? (Descriptive, diagnostic, predictive, or prescriptive)?

Answer 4

Prescriptive (“how do I make it happen?”) — NLP enables systems to interpret and act on human language. For example, using NLP to extract invoice data and automatically match it against a listing — automating an action rather than just analyzing.

Question 5

What type of analytic is drill-down and drill-through analysis? (Descriptive, diagnostic, predictive, or prescriptive)?

Answer 5

Diagnostic (“why did it happen?”) — Drilling down means peeling back layers of aggregated data to find the root cause of something. For example, if the current ratio drops, you drill down into its components to find out why.

Question 6

What type of analytic is what-if analysis? (Descriptive, diagnostic, predictive, or prescriptive)?

Answer 6

Prescriptive (“how do I make it happen?”) — What-if analysis tests different scenarios and answers the question “what scenario produces the best outcome?” It prescribes decisions rather than just describing or predicting.

Question 7

What type of analytic is forecasting? (Descriptive, diagnostic, predictive, or prescriptive)?

Answer 7

Predictive (“what will happen?”) — Forecasting uses historical trends and patterns to project future values — it answers “what will happen?” rather than “what happened?”

Question 8

What type of analytic is regression analysis? (Descriptive, diagnostic, predictive, or prescriptive)?

Answer 8

Predictive (“what will happen?”) — Regression uses historical data to model the relationship between variables and predict a dependent variable. For example: using hours worked and revenue to estimate supplies expense.

Question 9

What type of analytic is data profiling? (Descriptive, diagnostic, predictive, or prescriptive)?

Answer 9

Diagnostic (“why did it happen?”) — Data profiling examines the structure, completeness, and patterns within a dataset to understand the nature and quality of the data and explain anomalies. It asks “what patterns might explain this?”

Question 10

What type of analytic is data mining and discovery? (Descriptive, diagnostic, predictive, or prescriptive)?

Answer 10

Diagnostic (“why did it happen?”) — Data mining scans large datasets to uncover hidden patterns, correlations, or relationships that explain why something happened in the data. You’re mining for the “why”.

Question 11

What type of analytic is sentiment analysis? (Descriptive, diagnostic, predictive, or prescriptive)?

Answer 11

Predictive (“what will happen?”) — Sentiment analysis processes text (like social media posts) to predict attitudes or future behaviors — e.g., negative product sentiment predicting abnormal returns.

Question 12

What type of analytic is classification? (Descriptive, diagnostic, predictive, or prescriptive)?

Answer 12

Predictive (“what will happen?”) — Classification uses historical data to assign new data points to a category or class. It predicts what will happen by asking “what usually happens to this category/class?” For example, predicting whether a company’s ratio profile resembles bankrupt vs. solvent companies.

NOTE: Binning is descriptive, because you’re simply “binning” what’s already there (like manually sorting emails into the correct folder). This is different from classification, which is predictive - classification categorizes future items based on current characteristics (like marking emails as spam, so that all future emails will be sorted the same way - you’re predicting that any future email from that address will also be spam).

Question 13

What type of analytic is period-over-period analysis? (Descriptive, diagnostic, predictive, or prescriptive)?

Answer 13

Diagnostic (“why did it happen?”) — Zooming in on the period in which a change occurred helps identify the “why” — for instance, pinpointing that a spike in Q3 was driven by a single large transaction.

Question 14

What type of analytic is variance analysis? (Descriptive, diagnostic, predictive, or prescriptive)?

Answer 14

Diagnostic (“why did it happen?”) — Variance analysis compares actual results to expected or budgeted results and investigates the underlying causes of differences — it’s focused on explaining why a gap exists.

Question 15

What type of analytic is a sequence check? (Descriptive, diagnostic, predictive, or prescriptive)?

Answer 15

Diagnostic (“why did it happen?”) — A sequence check looks for gaps or duplicates in prenumbered documents. Missing numbers signal can signal the “why” behind unusual balances (ex: control failures)

Question 16

What type of analytic is decision support and automation? (Descriptive, diagnostic, predictive, or prescriptive)?

Answer 16

Prescriptive (“how do I make it happen?”) — These tools actively recommend or automate decisions based on analytical outputs. The goal is to guide action toward a desired outcome, not just describe or forecast.

Question 17

What type of analytic is machine learning? (Descriptive, diagnostic, predictive, or prescriptive)?

Answer 17

Prescriptive (“how do I make it happen?”) — In its most advanced applications, machine learning continuously learns from new data and prescribes optimal actions. It goes beyond prediction to actively recommending what should be done.

Question 18

What type of analytic is time-series modeling? (Descriptive, diagnostic, predictive, or prescriptive)?

Answer 18

Predictive (“what will happen?”) — Time-series models analyze data points collected over time to identify trends and seasonal patterns, then use those patterns to predict future values.

Question 19

What type of analytic is summary statistics (mean, median, range, standard deviation)? (Descriptive, diagnostic, predictive, or prescriptive)?

Answer 19

Descriptive (“what happened?”) — Summary statistics describe the characteristics of a dataset — they tell you what the data looks like, not why it looks that way or what might happen next.

Question 20

What type of analytic is aging data / frequency distribution? (Descriptive, diagnostic, predictive, or prescriptive)?

Answer 20

Descriptive (“what happened?”) — Aging buckets receivables by how long they’ve been outstanding (e.g., 0–30, 31–60 days). It summarizes and describes the current state of the data without explaining causes or predicting outcomes.