Fraud detection tool identifying suspicious transactions to prevent revenue loss and improve operational efficiency. Analyzes transaction data to detect anomalies like low-value sales, excessive refunds, and employee discounts.
Technical Specs
Daily queries on the server use Laravel's Task Scheduling to analyze data for:
Open tickets: Issue-marked transactions by stores under the user's jurisdiction.
Low Dollar/Modified Transactions: Transactions with sales less than $0 or cancellations.
Giveaways: Transactions with 0% net sales.
Refunds: Voided transactions.
Employee discounts: Transactions with discounts for employees.
Repeat Offenders: Employees under the user's jurisdiction with repeated offenses.
User management database is centralized and separate from the main application, connected to two databases.
Internationalization (i18n) implemented.
Built as a Single Page Application (SPA).
Initially deployed on an Apache 2 physical server, later moved to Amazon EC2 with CentOS 7 OS.
RESTful API architecture.
State management with Vuex.
OAuth2 authentication via Laravel Passport.
Features
Dashboard Overview - Transactions that marked as issues that need resolution.
Tickets overview
View raw transaction details
History Data - Report for total number of open tickets per store and total dollar amount of the open ticket per store.
Arl Report Completion - Report for total number of open tickets per ARL, closed / non-issue, incidents, amount of closed and incidents, completion rate.
Competitive Report - Variation between previous and current sales
Reports overview
Ticket resolution
User Management
Store Assignment / User Hierarchy
Historical Report - History summary of tickets
Data Feeds Generate data transactions manually (Same method used on daily task scheduling).
Taco Sales Report - Specific sales report for taco bell
Audit Trail
Threshold Settings
Forgiveness Period
- Option to choose on deciding how to resolve an incident ticket.
Modified Transactions
Refund
Discount
Coupon
Date Range - Default date range of data to be pulled
Re-process - Generate missing data transactions
Store Directory - Grant user access to stores
With over thousands of data being pulled each modules, all data are properly indexed and rendered from server to avoid long time waiting.
Role
Contributions
Responsible for overall development including UI/UX design
Freestyle design, no wireframes
Worked closely with project manager, system analyst and system admin