StayScope

StayScope is a comprehensive asynchronous analytics platform designed to transform hotel booking data into actionable insights. It integrates advanced data processing and visualization tools to support decision-making in the hospitality and hotel industry.

DEMO can be found below.

Technologies and Tools

• Python
• FastAPI - REST API
• Elastic Search
• LLM - Ollama (Mistral)
• SQLAlchemy (ORM) - PostgreSQL
• Docker
• Streamlit
• Plotly

Structure

• data: Data - CSV files.
• scripts: Python scripts for processing, REST API and Elasticsearch.
• webapp: Frontend application with Streamlit.
• notebooks: Jupyter notebooks for exploratory data analysis (EDA).

---

1. Main Application File (main.py): Entry point to your FastAPI application.
2. Database Models (models.py): Contains SQLAlchemy models and Database operations.
3. Elasticsearch Operations (elasticsearch_operations.py): Handles Elasticsearch indexing, searching, and other operations.
4. Schemas (schemas.py): Pydantic models for request and response validation.
5. API Routes (api_routes.py): FastAPI route definitions for Elasticsearch operations.
6. LLM (llm_model.py): Contains LLM model for report interpretation.
7. ETL (etl_utils.py): Incorporates an Extract-Transform-Load (ETL) approach for data processing.
8. Configurations (config.py): Configuration settings (database URLs, Elasticsearch connection details, etc.).
9. Visualization (fetch_utils.py): Visualization functions for fetched data from REST API.
10. User Interface - Frontend (streamlit_webapp.py): User Interface (Frontend) for interaction.

---

Installation

1. Clone the repository

git clone https://github.com/nedimcanulusoy/StayScope.git
cd StayScope

2. Create venv and install requirements.txt

python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

3. Set up elastic search via docker

docker-compose build

and

docker-compose up -d

NOTE: I assume that you have PostgreSQL running on your computer or server, either locally or through Docker. If you don't, please set one up before proceeding to the next step.

4. Pull LLM model for Ollama

NOTE: Default model is Mistral and feel free to change with anything you want to use in config.py file. LLM model variable is called LLM_MODEL

ollama pull mistral

5. Run REST API and Streamlit User Interface

REST API:

cd src

uvicorn main:app --reload

Streamlit:

cd webapp

streamlit run streamlit_webapp.py

NOTE: You need to create a config.py file to be able to run the project. Here is a template for it;

DATABASE_URL = 'your-full-postgresql-url'
DATABASE_TABLE_NAME = 'your-database-table-name'
ELASTICSEARCH_SETTINGS = {
    'host': 'es-running-host',
    'port': 'es-running-port',
    'scheme': 'http', #default.
    'auth': ('user', 'secret') #default.
}
ES_INDEX_NAME = 'your-es-index-name'
DATA_PATH = 'your-data-file-path'
DATA_FOLDER_PATH = "your-data-folder-path"
TMP_PATH = "your-tmp-data-folder-path"
TMP_CSV_FILENAME = 'your-tmp-data-file-name'

LLM_MODEL = ['mistral','gemma:7b', 'gemma:2b','phi']
PROMPT_TEMPLATE = """
your fancy prompt
"""

#ENDPOINT LIST, SUGGESTED TO KEEP AS IT IS.
ENDPOINTS = {
        "top_countries": "/api/v1/reports/top_countries",
        "cancellation_rate": "/api/v1/reports/cancellation_rate",
        "adr_by_month": "/api/v1/reports/adr_by_month",
        "length_of_stay_distribution": "/api/v1/reports/length_of_stay_distribution",
        "special_requests_impact_on_cancellations": "/api/v1/reports/special_requests_impact_on_cancellations",
        "average_lead_time_by_cancellation_status": "/api/v1/reports/average_lead_time_by_cancellation_status",
        "bookings_distribution_by_room_type": "/api/v1/reports/bookings_distribution_by_room_type",
        "bookings_by_guest_country": "/api/v1/reports/bookings_by_guest_country",
        "booking_source_analysis": "/api/v1/reports/booking_source_analysis",
        "booking_trends_over_time": "/api/v1/reports/booking_trends_over_time",
        "revenue_analysis_by_room_and_month": "/api/v1/reports/revenue_analysis_by_room_and_month",
        "impact_of_lead_time_on_adr": "/api/v1/reports/impact_of_lead_time_on_adr",
        "analyze_booking_composition": "/api/v1/reports/analyze_booking_composition",
        "correlate_cancelations_with_factors": "/api/v1/reports/correlate_cancelations_with_factors",
        "correlate_adr_with_factors": "/api/v1/reports/correlate_adr_with_factors",
        "analyze_repeat_guest_bookings": "/api/v1/reports/analyze_repeat_guest_bookings",
}

---

Demo

demo.mp4