AKbot

This project was developed as the final project for the 7-week Bootcamp jointly organized by Akbank and Patika.

Features:

• RAG: Provides answers to all your Akbank questions
• PandasAI: Delivers the analysis you need from your spending history
• Machine Learning: Predicts next month's prices by category
• Campaign Recommendations: Combines GenAI and rule-based systems to suggest personalized campaigns

Technologies Used

• Frontend: HTML, CSS, JavaScript, Streamlit
• GenAI: LangChain, OpenAI, PandasAI
• Machine Learning: scikit-learn, XGboost
• Database: SQLite, Chroma

Architecture

---

Requirements

Environment

Ensure that your Python version is set to 3.10.12 (pip version is 24.1.2):

python --version

• Setting up Virtualenv:

pip install virtualenv

• Creating a Virtual Environment:

virtualenv venv

• Activating the Virtual Environment:

source venv/bin/activate

• Installing the necessary libraries:

pip install -r requirements.txt

Configuration

• Set up your .env file:

cd <project-directory>

- Create the .env file and add your OPENAI_API_KEY:

    OPENAI_API_KEY='key' # .env file

Create VectorDB

python3 create_database.py

Create ML Model

python3 model.py

Run

• Launch the Streamlit app in terminal:

streamlit run akbot_streamlit.py

---

akbotdemo.mp4