|
1 | | -# CS203-Project |
| 1 | +# Trade Optimization Pathfinder (TOP) |
| 2 | + |
| 3 | +*An all in one Tariff Management Platform to streamline the process of calculating Tariffs and optimizing the trade routes between 2 or more countries* |
| 4 | + |
| 5 | +## 🌍 Problem Statement |
| 6 | +Design and implement a system called TARIFF (Trade Agreements Regulating Imports and Foreign Fees): its purpose is to define import tariffs (and/or additional fees) and to calculate them against a certain industry or product between different countries at any given times. |
| 7 | + |
| 8 | +## 💡 Our Solution |
| 9 | +**Trade Optimization Pathfinder (TOP)** is the perfect solution to tackle Trade Route Tariff Calculations and Management through our all in one optimizer platform that: |
| 10 | +* Provides clear **Trade Route Tariff Calculations**. |
| 11 | +* Enables admins to **manage Tariffs**. |
| 12 | +* Gives **route visualization** a bird’s-eye view for smarter coordination. |
| 13 | + |
| 14 | +## 🎯 Features |
| 15 | + |
| 16 | +### Baseline Tariff Calculation Process |
| 17 | +<img width="1089" height="1293" alt="image" src="https://github.com/user-attachments/assets/3a2636c5-0ebe-4432-859d-668e9f5abbe0" /> |
| 18 | +* Step 1: Enter source and destination countries to search for goods using HSCode or descriptions |
| 19 | +<img width="1174" height="1240" alt="image" src="https://github.com/user-attachments/assets/5a4de1f6-6702-415f-a030-c35f7dbdfe7e" /> |
| 20 | +* Step 2: System returns the tariff(s) available and user selects from the dropdown and enters quantity of product |
| 21 | +* Step 3: Select desired currency conversion and calculate the result |
| 22 | + |
| 23 | +## Exchange Rate Conversion |
| 24 | +Exchange rates are from ExchangeRateAPI, pulled every hour and stored in Redis, ensuring sped and scaleability without sacrificing accuracy in exchange rate conversions. |
| 25 | + |
| 26 | +## Explainable Calculation |
| 27 | +TOP breaks down calculation into step by step formula and displaying each step |
| 28 | +This not only gives users confidence about the correctness of the final tariff calculation but also help beginners learn about different tariffs and how they are calculated |
| 29 | + |
| 30 | +## Compliance Checklist |
| 31 | +<img width="969" height="791" alt="image" src="https://github.com/user-attachments/assets/b454477a-000a-4f65-84b5-25b78d1360fb" /> |
| 32 | +Offers users a list of checklist items to look out for after they perform a calculation. This includes items like Compliance declarations and Certification and classification requirements. |
| 33 | +Uses Pinecone and OpenAI for semantic search, with strict JSON output for frontend checklist visualisation. |
| 34 | +Checklist items are sourced from official reports of government websites and offer a direct source link from our app for users to verify the authenticity of the checklist item. |
| 35 | + |
| 36 | +## Route Visualization |
| 37 | +<img width="1134" height="380" alt="image" src="https://github.com/user-attachments/assets/edd0eb3e-d2e7-4df7-a15d-7bd8c49e0870" /> |
| 38 | +Gives 4 distinct routes: Each optimized for a certain aspect: |
| 39 | + |
| 40 | +*Risk: Risk of piracy/incidents occuring |
| 41 | +*Cost: Cheapest way to ship objects |
| 42 | +*Carbon Emissions: Most environmentally friendly route |
| 43 | +*Time: Fastest way to transport goods |
| 44 | + |
| 45 | +Each route also gives a breakdown of all 4 factors for more in depth comparisons |
| 46 | + |
| 47 | +### AI Chatbot |
| 48 | +<img width="1141" height="456" alt="image" src="https://github.com/user-attachments/assets/8c4d19e7-ce33-419b-b960-98e8be3264d0" /> |
| 49 | +Our chatbot uses a RAG system, when a user asks a question, the system doesn’t just generate an answer from the language model alone. Instead, it first checks our Pinecone vector database that stores reliable tariff data. |
| 50 | + |
| 51 | +Pinecone then returns the most relevant context back to GPT - for example, specific tariff definitions, regulatory notes, or documentation references. |
| 52 | + |
| 53 | +Using these together, GPT generates a final answer that is grounded in real data, rather than guessing. This ensures the chatbot’s responses are accurate, reliable, and source-backed. |
| 54 | + |
| 55 | +<img width="869" height="1221" alt="image" src="https://github.com/user-attachments/assets/c1b91c86-e918-4ec3-989a-4213dac912c9" /> |
| 56 | +Administrators can add new documents or update existing tariff information directly into Pinecone. |
| 57 | + |
| 58 | +### Results Tab |
| 59 | +This tab allows users to quickly build new routes, compare Tariffs and export results as a PDF |
| 60 | + |
| 61 | +## Multi-Leg Route Builder |
| 62 | +<img width="827" height="609" alt="image" src="https://github.com/user-attachments/assets/5022f312-e24d-486f-b701-8c9741d8bbbc" /> |
| 63 | + |
| 64 | +Users have the ability to combine multiple routes through multiple countries together for further cost optimization, to allow users to compare metrics for direct shipping versus multi-leg shipping routes as well. |
| 65 | + |
| 66 | +## Tariff comparator |
| 67 | +<img width="1077" height="403" alt="image" src="https://github.com/user-attachments/assets/2b5c1a3a-c13e-4b7c-aa15-c61be1e0bb38" /> |
| 68 | +Users are able to compare different tariffs between two different trade routes, and gives a breakdown on which is more optimal based on one of the 4 metrics: |
| 69 | +*Risk: Risk of piracy/incidents occuring |
| 70 | +*Cost: Cheapest way to ship objects |
| 71 | +*Carbon Emissions: Most environmentally friendly route |
| 72 | +*Time: Fastest way to transport goods |
| 73 | +We also provide some additional analytics as shown: |
| 74 | +<img width="761" height="97" alt="image" src="https://github.com/user-attachments/assets/23fe14a4-9315-4119-873e-dc0221e57612" /> |
| 75 | + |
| 76 | + |
| 77 | +### Security |
| 78 | +<img width="904" height="630" alt="image" src="https://github.com/user-attachments/assets/3a191a7b-7f74-4539-8d99-14e1ac63aa24" /> |
| 79 | +Secured by AWS cognito, ensuring compliance with strict regulations and auditability as required by banks. |
| 80 | +Designed with strict least privilege access principles, with 2 distinct user groups: Users and Admins. |
| 81 | +## Users: |
| 82 | +Allowed to use the Tariff Calculation Features, Results tab and AI Chatbot |
| 83 | +## Admins: |
| 84 | +Are allowed to use anything the Users have accessed to, as well as Tariff Management page |
| 85 | + |
| 86 | +## Tariff Management |
| 87 | +A page that only Admins are able to access |
| 88 | +It has the following functionalities: |
| 89 | +*Add Tariff/Delete Tariff: |
| 90 | +Due to frequently changing regulations, admins have the authority to update the system to stay compliant even outside of the regularly updated intervals. |
| 91 | + |
| 92 | +*Edit Tariff: |
| 93 | +Any mistakes in tariff entries can be rectified without having to recreate the entire database and without developer intervention |
| 94 | + |
| 95 | +*Adding Information to Chatbot: |
| 96 | +Able to modify the compliance documents that is used as context for the AI chatbot |
| 97 | + |
| 98 | + |
| 99 | + |
| 100 | +## 🛠️ Technical Implementation |
| 101 | + |
| 102 | +* **Frontend:** React, TypeScript (map visualisation via Mapbox). |
| 103 | +* **Backend:** Java (Spring Boot), Python. |
| 104 | +* **Databases:** PostgreSQL, Neo4J, Pinecone, Redis. |
| 105 | +* **Deployment:** Vercel (Frontend), AWS EC2 (Backend). |
| 106 | + |
| 107 | + |
| 108 | +# To Run Locally |
2 | 109 |
|
3 | 110 | ## 1. Project Setup Instructions |
4 | 111 |
|
|
0 commit comments