Back to Projects
Data Engineering
ELT
YOLOv8
Dagster

Telegram Medical Data Pipeline

End-to-end ELT pipeline for healthcare analytics

Project Overview

An end-to-end data pipeline to extract, transform, and analyze data from public Telegram channels focusing on Ethiopian medical businesses. The pipeline provides actionable insights into medical product trends, pricing, and visual content.

Business Context

Kara Solutions aims to build a robust data platform to generate insights about Ethiopian medical businesses. This platform leverages data scraped from public Telegram channels to answer critical business questions and enhance data analysis capabilities.

Key Features

Automated Data Scraping: Extracts messages and images from specified Telegram channels

Robust Data Lake: Stores raw, immutable data for auditing and re-processing

Dimensional Data Warehouse: Implements star schema in PostgreSQL for efficient analytical queries

AI-driven Data Enrichment: Utilizes YOLOv8 for object detection on images

Pipeline Orchestration: Manages entire workflow using Dagster for reliability and monitoring

Technical Implementation

ELT framework: Extract raw data, load into data lake, transform using dbt

Telegram API integration using Telethon for message and media extraction

PostgreSQL data warehouse with dimensional modeling

Computer vision integration with YOLOv8 for image analysis

FastAPI analytical endpoints for structured data access

Technologies Used

Dagster
PostgreSQL
dbt
YOLOv8
FastAPI
Telethon
Docker
Docker Compose

Results & Impact

Enhanced healthcare analytics through efficient data pipeline management.

Interested in Learning More?

Explore the complete implementation, documentation, and code on GitHub.