Introduction to Data Analytics
A visual, interactive overview of the modern data ecosystem, analytics process, data types, repositories, ETL, wrangling, statistics, mining, and visualization, with accounting examples, Alteryx workflows, AI/agentic AI discussion, and interactive checks for ACCT 311.
Accounting data analytics turns accounting records into insight.
In ACCT 311, you will use analytics to ask better accounting questions, prepare decision-ready data, and communicate what the numbers mean for performance, risk, and operations.
Analyze monthly revenue, gross margin, and seasonality to explain what changed and why.
Identify overdue customers, estimate collection risk, and support cash-flow planning.
Track inventory movement, slow-moving items, and relationships between units, cost, and sales.
Spot unusual journal entries, duplicate payments, or inconsistent transactions requiring review.
Data now comes from everywhere.
Organizations work with structured and unstructured datasets from text, images, video, click streams, conversations, IoT devices, social media, real-time events, legacy databases, and professional data providers.
More Variety
Tables, files, sensor feeds, documents, images, and conversations can all become analytical inputs.
More Velocity
Data can arrive continuously, requiring faster cleaning, storage, and interpretation.
What does a data analyst actually do?
A data analyst translates data and numbers into plain language so organizations can make decisions.
Look for completeness, quality, and consistency.
Click to flipCan this dataset answer the business question responsibly?
Fix missing values, duplicates, outliers, and inconsistent formats.
Click to flipWhat needs to be corrected before analysis begins?
Visualize and communicate findings for decision makers.
Click to flipWhat should the audience understand and do next?
Three forces changed the data ecosystem.
Cloud computing, machine learning, and big data provide access to scalable storage, powerful computing, and advanced analytical tools.
Access resources without owning all infrastructure.
FlipLarger datasets can be stored, shared, and processed more easily.
Models can learn patterns from data.
FlipAnalysts can use predictions, classifications, and anomaly detection.
Datasets can be too large or complex for traditional tools.
FlipDistributed storage and processing become more important.
Better tools make faster analysis possible.
FlipThe analyst still needs judgment, context, and communication.
Data analytics is a process, not just a tool.
It includes gathering, cleaning, analyzing, mining, interpreting, and reporting data so findings can be used.
Each type asks a different question.
Summarize past or current results.
FlipGross margin decreased by 2 points this quarter.
Investigate causes, relationships, and drivers.
FlipReceivables increased because a few major customers delayed payment.
Estimate likely outcomes using data and models.
FlipCash collections are expected to improve next month based on recent trends.
Recommend actions based on analysis.
FlipTighten credit review and follow-up in the highest-risk customer segment.
Good analytics starts with a clear problem.
The process moves from a decision need to a meaningful, communicated finding.
- Understand the problem and desired result
- Set a clear metric
- Gather and clean data
- Analyze, mine, and interpret results
- Communicate findings clearly
Put the analytics process in order.
Drag the cards into the most logical sequence, then check your answer.
Related, but not identical.
Detailed Examination
Data analysis examines elements, structures, or patterns. It often focuses on what can be inferred from a specific dataset or situation.
- Can be broader than numerical data
- Often focused on examination
- Can support one part of analytics
Systematic Data Process
Data analytics usually implies computational analysis of data or statistics and can include predictive or action-oriented elements.
- Usually uses data for inference
- Includes end-to-end workflow
- Supports decision-making
Analysts connect data work to real organizational decisions.
Responsibilities
- Extract transaction-level data from accounting systems and supporting files
- Clean, reconcile, standardize, and document accounting data
- Test trends, ratios, exceptions, and relationships across periods
- Use visuals and summaries to explain performance, risk, and control issues
- Create reproducible workflows for review, audit, and reporting
Accounting Applications
- Financial reporting and financial statement articulation
- Receivables aging, inventory analysis, and cash-flow review
- Budget versus actual variance analysis
- Employee expense, procurement, or duplicate-payment anomaly checks
- Management dashboards for sales, margin, and operating performance
Example: employee expense, procurement, or duplicate-payment anomaly checks
Here is how the data analytics process can be applied to a practical accounting control problem. We will use duplicate-payment anomaly checks as the worked example, but the same logic also applies to employee expense and procurement review.
How the Process Works
- Define the question: Are we paying the same invoice more than once?
- Gather data: AP invoices, payment file, vendor master, PO or expense records.
- Clean the data: standardize invoice numbers, vendor names, dates, and amounts.
- Analyze for anomalies: same vendor + same amount + same invoice number, or near-matches with date differences.
- Interpret and communicate: separate likely duplicates from valid split payments or credit adjustments.
What Might Be Flagged
- Exact duplicate invoice number and amount
- Same vendor and amount paid twice within a short period
- Negative balances or reversals that do not make sense
- Payments without matching invoice support
- Vendors with unusually high exception counts
Strong analysts combine technical, functional, and soft skills.
Spreadsheets, statistical tools, visualization tools, programming, queries, and repositories.
FlipExcel, Power Query, Alteryx, Power BI, SQL basics, and repeatable analytics workflows.
Statistics, analytical techniques, visualization, and project management.
FlipChoosing the right metric, method, visual, and interpretation.
Collaboration, data storytelling, stakeholder engagement, curiosity, and intuition.
FlipExplaining findings in clear language for decision makers.
Data differs by structure and analysis approach.
Analysts need to know both where data came from and what format it is in.
The source tells you how the data was created and what controls may exist. The format tells you how the data is stored and what preparation may be required. Two files can contain the same accounting information but require very different workflows.
File Formats
CSV/TSV are simple delimited tables. XLSX can contain multiple sheets, formulas, formatting, and hidden assumptions. JSON/XML are common in APIs and system integrations. PDF is presentation-oriented, so tables often need extraction and reconciliation before analysis.
Internal Accounting Systems
ERP systems, general ledger detail, accounts receivable and payable subledgers, payroll, purchasing, inventory, POS systems, bank feeds, customer/vendor master files, and approval logs. These often contain the transaction-level evidence used for accounting analysis.
External Sources
SEC filings and XBRL, market data, economic indicators, customer or vendor data, benchmarks, credit information, and public APIs. External data can enrich analysis but may use different definitions, periods, or levels of aggregation.
Continuous or Event Data
Web orders, payment confirmations, inventory scans, approval timestamps, card transactions, and IoT events can arrive continuously. Ask whether the stream is complete, whether late records exist, and which timestamp determines the reporting period.
If the same monthly sales total appears in an ERP export, a manager's Excel workbook, and a PDF report, which source would you trust first? What evidence would you use to decide?
Accounting Source Examples
Mystery files: can you identify the format?
Look at how the information is structured and what an analyst would need to do with it.
Which of these formats would you least want to receive for a 500,000-row transaction analysis, and why?
ETL turns raw data into analysis-ready data.
Extract
Collect data from source locations such as databases, files, systems, APIs, or providers.
Transform
Clean, enrich, standardize, validate, and reshape data for analysis.
Load
Place processed data into a destination repository, application, or analytics system.
Alteryx shows the workflow directly on the canvas.
Instead of just describing the process, Alteryx lays out the data-preparation logic visually. That makes recurring accounting work easier to understand, review, rerun, and improve.
What It Does
Imports invoices and payments, joins customer credit terms, calculates days outstanding, assigns aging buckets, and prepares a clean AR-aging dataset.
Why It Matters
The same documented workflow can be rerun next month on a new file instead of rebuilding the cleaning steps manually.
What Students Should Notice
The canvas shows tool order, joins, formulas, and checks visually. That is a major difference from writing hidden transformation logic somewhere else.
Wrangling is iterative: explore, transform, validate.
Wrangling prepares messy data for analysis by manipulating, combining, normalizing, denormalizing, cleaning, and enriching datasets.
Bring related datasets together.
FlipBad joins can duplicate rows or drop records.
Missing values, duplicates, errors, formats, and outliers.
FlipCleaning choices can change the conclusion.
Statistics helps analysts understand what the data represents.
Descriptive Statistics
Summarize what the data shows.
- Central tendency
- Dispersion
- Skewness
Inferential Statistics
Make generalizations or inferences from data.
- Hypothesis testing
- Confidence intervals
- Regression analysis
Insights become useful when patterns are communicated clearly.
Data Mining
Extract knowledge by identifying correlations, patterns, variations, and trends using pattern recognition, statistics, and mathematical techniques.
Data Visualization
Communicate information through visual elements such as graphs, charts, and maps so findings are easier to understand, interpret, and remember.
Generative AI can assist analysis. It does not remove the need to understand the data.
AI can help explain formulas, draft SQL or Alteryx logic, suggest data-quality checks, interpret a chart, or propose a workflow. The accountant still needs to verify the source, logic, totals, assumptions, and conclusion.
You Direct Each Step
- Suggest an Excel formula
- Explain an Alteryx Join tool
- Draft SQL or DAX
- Recommend a visualization
- Help interpret an unusual trend
The System Can Pursue a Goal
- Plan a sequence of tasks
- Choose among available tools
- Inspect outputs and feedback
- Adapt the next action
- Stop or escalate when criteria are met
Example AI Prompt
“I have invoice, payment, and customer tables. Help me identify the fields needed for AR aging, propose validation checks, and explain how I could build the workflow in Alteryx.”
Do Not Skip Verification
AI can confidently use the wrong field, misread a business definition, produce a bad join, or invent an explanation. Reconcile totals and inspect exceptions before accepting the result.
If AI produces the correct answer but you cannot explain how it got there, is that acceptable accounting work? Why or why not?
Think of agentic AI as a small team of narrow helpers around a workflow.
The four items below are not meant to replace Alteryx tools. They are examples of specialized AI helpers that could support the same AR-aging process. Each one has a narrow job, passes results to the next step, and stops when a control rule fails.
Before anything else, it checks whether invoice date, customer ID, amount, payment amount, and due-date fields are present and in the right format.
After invoices, payments, and customer terms are joined, it reports matched and unmatched rows and checks whether record counts or totals changed unexpectedly.
It calculates days outstanding, assigns aging buckets, and flags duplicates, credits, unexpected negative balances, or overdue items that need review.
Only after the checks pass, it writes a plain-language summary of the verified output for the accountant or manager without changing source data.
How This Fits
Alteryx is the visible workflow canvas. Agentic AI would sit around that process to check inputs, inspect joins, flag issues, and explain verified results. In other words, it supports the workflow rather than replacing accounting judgment.
Example Stop / Escalation Rule
“If unmatched invoice value exceeds 1% of total receivables, or if negative balances appear in unexpected accounts, stop and request human review before producing a final aging report.”
Check your understanding.
Your Quiz Score
Use this score as a quick self-check. Revisit any slide where the feedback surprised you.
Your Turn: Complete the Excel Skills Check.
Finish this introduction by working directly with an accounting transaction dataset. This is a low-stakes diagnostic, not an Excel exam. Use what you know, try the stretch tasks if you are ready, and show your current level of comfort with Excel.
Core Tasks
- Format the data as a table
- Sort and filter transaction rows
- Calculate totals with basic formulas
- Identify overdue and high-value items
Stretch Tasks
- Use XLOOKUP to bring in customer information
- Create a PivotTable summary
- Flag possible data-quality issues
- Write a short reflection on your Excel comfort level
Save Your Work
Save the completed workbook using a recognizable file name such as LastName_FirstName_ExcelSkills.xlsx.
Submit on Canvas
When you finish, upload your completed Excel workbook to the designated Canvas assignment. Submit the actual .xlsx file, not a screenshot or PDF.