This project offers a python solution for risk assignment in financial datasets, often applied in auditing processes. In addition to the main Risk_Rating_Assignment file, the auxiliary processor file preprocesses the data, including clustering similar invoice descriptions using the DBSCAN clustering algorithm.
The potential risks this program can identify are as follows:
- a - accounting code is missing or incorrect
- b - the events 1 through 15 above happen out of order (ex: invoice date is prior to purchase order date)
- c - bill of sale or delivery documents are missing or do not agree with amounts on invoice
- d - bill of sale or delivery documents are missing or do not agree with purchase order
- e - purchase order quantity does not agree with invoice quantity
- g - control documents or elements of the audit trail are missing: purchase order, bill of sale, invoice, accounting code
- h - the audit trail of transactions within a Department (Marketing, IT, U.M.) are inconsistent
- i - duplicate payments – same payments to same vendor for same good/service or same payment for same service/good to two or more different vendors
- j - payments 2% under the threshold for competitive bidding
- k - payments 2% under the threshold of higher or Chief approval
- l - payments to a vendor for the same service/good that accumulate to be in excess of blanket purchase order or contract
- m - split purchase order - multiple POs to same vendor for same items
- n - split payments - payments to same vendor for same item and same date
original dataset not included for privacy reasons