Open Data Stories

Why?

Most important information still lives in dashboards, spreadsheets, PDFs, and websites people rarely check, while short-form video has become one of the main ways people consume content online.

This project explores programmable media: turning live data streams into automated localized content for Moroccan audiences.

Unlike most AI-generated content, the system is deterministic and data-driven. The output is generated from structured data and predefined logic, making it predictable, reproducible, and easier to verify.


What is it?

I build automated TikTok videos that transform real-world data into short-form visual content.

Current videos include:

  • Moroccan currency exchange rates
  • Casablanca Stock Exchange summaries
  • Daily weather forecasts

All videos are generated and published automatically on a schedule.


How does it work?

The system collects data from APIs and scraping pipelines, processes it with Python, then generates fully rendered videos programmatically using Remotion.

The entire system runs on a small VPS with 4GB RAM alongside other projects, with Telegram notifications for data collection, rendering status, and publishing events.

The project combines backend engineering, automation, and data visualization into a scalable media-generation pipeline.

Check it out

Stack

PythonFastAPIBeautifulSoupJavascriptRemotionnginxSQLitetelegram

Phases

1

Phase 1done

Testing the waters

Build initial video templates for different content types (weather, stock market, currency rates).

javascripttailwindremotion

Explore what is possible with the Remotion framework and validate the idea of programmatic video generation.

2

Phase 2done

Automation first steps

Start building data retrieval and cleaning scripts for all content sources.

pythonBeautifulSouppdfplumber

Move from static videos to real-time, data-driven content pipelines.

3

Phase 3In progress

First full pipeline

Connect data ingestion, processing and video rendering into a single automated workflow.

pythonFastAPISQLiteTelegramnginx

Prove that end-to-end automated video generation is reliable and repeatable.

4

Phase 4planned

Observability & infrastructure

Add monitoring, metrics collection, and deployment automation to the pipeline.

PrometheusGrafanaAnsible

Treat the project like a real production system: observable, reproducible, and easier to maintain as automation grows.