| Course | ACC 427 Investigating with Computers |
|---|---|
| Module | Module 4 |
| Paper type | undergraduate data analytics project joining transaction and location data |
| Length | About 1,010 words, 6 pages |
| Format | APA 7 student paper |
| School | Southern New Hampshire University |
| Program | BS Accounting |
| Updated | October 2026 |
Free sample paper for ACC 427 Module 4
Where Was the Truck? Joining Fuel Card Transactions With GPS Data at a Composite Waste Hauler
[Student Name]
Southern New Hampshire University
ACC 427: Investigating with Computers
Project One
[Instructor Name]
[Date]
The organization, setting and figures below are a composite written as a model document. No real employer, client, colleague or patient is described.
Where Was the Truck? Joining Fuel Card Transactions With GPS Data at a Composite Waste Hauler
The Risk and the Data
Each of the waste hauler's 230 trucks has a fuel card, and supervisors and mechanics hold cards for service trucks and loaners. Fuel is the company's second-largest operating cost, about $9.8 million a year. The risk is that cardholders buy fuel for personal vehicles or for resale. The card provider's data show each purchase's card, station, time, gallons and amount; the trucks' telematics units record location every two minutes. Neither source alone shows misuse, because a purchase looks normal on the card statement and the GPS shows only where the truck went. Joined together, they show whether the truck was at the pump (Nigrini, 2020).
Acquisition and Validation
Twelve months of card transactions, 61,800 records, were obtained from the card provider with a monthly statement total, and the export reconciled to the statements within rounding. GPS data, 31 million pings, came from the telematics vendor with record counts by truck and day. Validation found 14 truck-days with no GPS data, which were listed so that purchases on those days would not be misread as distant.
The Join
Each card is assigned to a truck in the fleet system. For each purchase, the join found the assigned truck's GPS pings within 30 minutes before and after the purchase time and calculated the closest distance between the truck and the station. Purchases with no truck assignment or no pings in the window were kept in a separate list rather than dropped, so they could be reviewed rather than silently lost; 1,900 purchases, about 3 percent, fell into that list.
Tests and Thresholds
Table 1. Tests and Thresholds
| Test | Logic | Threshold | Reason for threshold |
|---|---|---|---|
| Distance | Truck far from station at time of purchase | More than 20 miles within 30 minutes | Trucks average 30 mph on routes; 20 miles allows for GPS error and timing |
| Tank capacity | More fuel than the truck can hold | Gallons above tank size | Tank sizes from fleet records |
| Off-hours | Purchase when truck is parked at yard | Outside scheduled route times and truck stationary | Route schedules from dispatch |
Results and Review
Table 2. Distance Test Funnel
| Stage | Count |
|---|---|
| Purchases tested | 61,800 |
| Matched to GPS | 59,900 |
| Flagged, truck more than 20 miles away | 512 |
| Cleared: GPS outage on that truck-day | 210 |
| Cleared: rental truck using a loaner card, documented by fleet office | 70 |
| Cleared: data errors, station coordinates wrong in provider data | 46 |
| Unexplained | 186 |
All 186 unexplained purchases were made on one card, assigned to a fleet supervisor's service pickup. They total 9,400 gallons and $33,600 over twelve months, mostly diesel at two stations near the supervisor's home, usually on weekends when the pickup was parked at the yard. The tank test flagged 41 of the same purchases as exceeding the pickup's 34-gallon tank. The off-hours test independently flagged 152 of them. The overlap of three independent tests on one card makes an innocent explanation unlikely, though it remains possible.
The Unmatched List
The 1,900 purchases that could not be matched to GPS deserved the same attention as the flags, because a cardholder who knew about telematics might use a card assigned to a truck without a working unit. Review found that 1,412 were on cards for loaner and rental vehicles without telematics, 371 fell on the 14 truck-days with no GPS data and 117 were on cards whose truck assignment had not been updated after trucks were replaced. None of the 117 showed unusual volume or timing, and the fleet office corrected the assignments. Had the join dropped these records silently, the project could not have claimed that the whole population was tested, and a second scheme could have hidden among them.
Why This Works and What It Cannot Show
Bierstaker et al. (2006) found that accountants regard data analysis as effective but use it less than they should; this project shows why it is worth the effort: a statement review would have seen ordinary purchases. Research on continuous data-level auditing shows how tests like these can run automatically against each day's data and route exceptions to a reviewer as they occur (Kogan et al., 2014), an arrangement that would likely have stopped this pattern within its first month or two. The analysis does not show who used the card or what vehicle received the fuel. The supervisor may have an explanation, such as fueling a company generator, which an interview must test.
Thresholds and Sensitivity
The 20-mile threshold was chosen from the fleet's average route speed and the 30-minute window, but a different cutoff would change the number of flags. At 10 miles, the distance test flags 1,140 purchases, many of them trucks fueling at stations just off their routes, which adds review work without adding findings; all 186 supervisor purchases are still flagged. At 40 miles, flags fall to 301, and 171 of the supervisor's purchases remain. The pattern on his card is robust to the threshold, which strengthens confidence that it is not an artifact of the cutoff. Reporting this sensitivity in the project answers the obvious question a reviewer would ask and shows that the result does not depend on a convenient choice.
Recommendations
The company should refer the 186 purchases to counsel and plan an interview with the supervisor after securing station receipts and any available station video. It should assign cards to vehicles rather than people, require odometer and vehicle number at the pump, cap gallons at tank capacity and run the distance test monthly with exceptions sent to the controller.
Conclusion
Joining 61,800 fuel transactions with truck GPS data and reviewing 512 distance flags left 186 unexplained purchases, all on one supervisor's card, totaling $33,600. Tank and timing tests independently support the pattern. The finding warrants an interview and immediate control changes, and the tests should be made continuous. The same join, run monthly against fresh card and telematics data, costs little once the scripts exist and would turn a twelve-month discovery into a one-month one, which is the practical return on building the analysis carefully the first time.
References
Bierstaker, J. L., Brody, R. G., & Pacini, C. (2006). Accountants' perceptions regarding fraud detection and prevention methods. Managerial Auditing Journal, 21(5), 520-535. https://doi.org/10.1108/02686900610667283
Kogan, A., Alles, M. G., Vasarhelyi, M. A., & Wu, J. (2014). Design and evaluation of a continuous data level auditing system. Auditing: A Journal of Practice & Theory, 33(4), 221-245. https://doi.org/10.2308/ajpt-50844
Nigrini, M. J. (2020). Forensic analytics: Methods and techniques for forensic accounting investigations (2nd ed.). Wiley.
What the ACC 427 Module 4 instructions ask for
Project One in ACC 427 usually asks you to design and carry out a data analytics plan for a specific fraud risk and report the results. Expect to define the risk and the data needed, describe how the data were obtained and validated, design tests with stated logic and thresholds, join tables or systems where useful, run the tests, review the flags and separate false positives from exceptions that need investigation, and recommend follow-up and controls. Present results with counts at each stage, from records tested to flags to confirmed exceptions, and explain how thresholds were chosen. The report should be clear enough for a manager to act on and detailed enough for another analyst to repeat.
How this ACC 427 Module 4 project one example is built
The project examines fuel card misuse. Fuel transactions from the card provider are joined to GPS pings from the trucks' telematics units by card, truck and time. A distance test flags purchases where the assigned truck was more than 20 miles from the station within 30 minutes of the purchase. A tank test flags purchases larger than the truck's tank. An off-hours test flags purchases outside scheduled routes. The distance test produces 512 flags; review clears 210 as GPS outages, 70 as rental trucks on borrowed cards and 46 as data errors. Of the remaining 186, all are on one fleet supervisor's card, totaling 9,400 gallons and $33,600. The report recommends controls and an interview.
Where the ACC 427 Module 4 rubric puts the points
Rubrics for ACC 427 Project One typically score the definition of the risk, data acquisition and validation, test design and thresholds, the join logic, the review of results, the recommendations and the clarity of the report. Top papers explain why each threshold was chosen, report counts at every stage, show how false positives were cleared and with what evidence, and connect the findings to specific controls. Graders reward joins that link systems in a way that exposes concealment, and a funnel table showing how flags were reduced to findings. Common deductions include thresholds chosen without reasons, flags reported as fraud without review, joins that drop unmatched records silently and recommendations unrelated to the findings.
ACC 427 Module 4 help: the mistakes that cost points
Analytics projects most often lose points when a join silently drops records, so that unmatched transactions, often the most interesting ones, never get tested. Another frequent problem is presenting every flag as a finding. If your project examines purchasing cards, travel expenses or payroll, the same structure of data, join, test, review and recommendation applies, and the plan can be built around your data. Report the funnel, records tested, flags raised, flags cleared, exceptions remaining, in a single table; it is the clearest way to show both rigor and judgment, and it answers the grader's first question. Then explain the threshold choices in a sentence each.
Get ACC 427 Module 4 written to your instructions
Send the ACC 427 Project One guidelines, data description and rubric. The report will state the risk, design tests with their logic and thresholds, describe joins and validation, report flags and how they were resolved, and recommend controls. Your first one costs nothing; allow about two days. The paper above is an original model document written by our desk, not a submitted student paper and not an official Southern New Hampshire University document.
More ACC 427 papers and related BS Accounting samples
- ACC 427 Module 1 Discussion: Where the Digital Evidence Lives
- ACC 427 Module 2 Data Acquisition Assignment: Getting a Complete Copy of the Scale Data
- ACC 427 Module 3 Data Analysis Assignment: Voids, Gaps and the 3 p.m. Pattern
- QSO 340 Module 8 Project Closure and Lessons Learned Report
- ACC 405 Module 4 Project One: A Consolidation Worksheet One Year After Acquisition
- ACC 318 Module 6 Discussion: A Covenant Breach and Where the Debt Goes
- BUS 210 Module 5 Project One Management Plan
ACC 427 Module 4 questions, answered
Where can I find a free ACC 427 Module 4 Project One sample?
This page includes a complete ACC 427 Module 4 Project One joining fuel card transactions with truck GPS data at a waste hauler.
What is a join in data analytics?
An operation that combines records from two tables using a shared key, such as card number and time, so related data can be tested together.
How are thresholds chosen for analytic tests?
From the data and the business, such as the distance a truck could travel in a given time or the tank's capacity, set so flags are meaningful but not overwhelming.
What is a false positive in fraud analytics?
A flagged record that turns out to be legitimate on review. Reporting how many flags were cleared, and why, is part of sound analysis.
What controls reduce fuel card misuse?
Assigning cards to vehicles rather than people, requiring odometer entry, limiting fuel types and amounts, and automated matching of purchases to vehicle location.