ACG 311 Final (Python) Project

Benford’s law give the expected patterns of digits in tabulated data. It is used in fraud analysis of financial statement accounts such as revenue, expenses, fixed assets, and earnings per share.

Write a program in python that perform the following actions:

Similar to the following article:

http://www.journalofaccountancy.com/issues/2003/aug/turnexcelintoafinancialsleuth.html

1. Take in more than 1.5 million rows from a csv formatted file. Use ten years of expenses of one agency from the OhioCheckbook.com to test or the num.txt file from the 2015q1 file at https://www.sec.gov/dera/data/financial-statement-data-sets.html.

2. Create plots of the benford frequency rate (0-9 and 0-100) and the actual frequencies of the: first digit, second digit, and first two digits from the data.

3. Output a file for each digit frequency that varies by more than one standard deviation from the benford less actual frequency statistics, for one year of expenses from one agency under the state of Ohio from ohiocheckbook.com that do not fall within the benford rate for further investigation. The user should choose the location of the output file (see Tkinter module asksaveasfilename). The output file should include the benford rate and the actual rate for the digits at the top of the sheet so the person investigating knows why the transactions are being investigated.

4. Document what inputs your programs needs, the outputs and why they are important. Documentation should be no longer than a single page single spaced 12 point font Times New Roman. The document also does not have be one page, it can be as short as you see fit. However, the document should be comprehensive enough that someone who knows nothing about python and benford’s law will be able to understand what the program does, why it is important and what the output shows.

Extra Credit:

10 point If the analysis, plot and viewing of the questionable transactions are done entirely in a GUI.

5 point if the GUI program is a Microsoft windows program (i.e. has a .exe extension).

5 points if the user can select the criteria used to determine which transactions to investigate further. Suggestions (mean, median, standard deviation 1, 2, or 3 percentiles, quartiles, specific number (single or double), etc.).