Filonov

When dealing with huge amounts of media (video, images, texts) it might be hard to identify which creative approaches work the best since actual performance is usually evaluated on a single medium.

Key features

• Fetching media: Downloading media info from various sources (i.e. Google Ads)
• Tagging media: Getting content of each media as a tag.
• Finding similar media: Find similar media and group them into clusters.
• Visualizing media: Ready to use dashboard for understanding media performance on tag, media, and cluster levels.

Installation

pipuv

pip install filonov

uv add filonov

Usage

filonov consists of two major parts:

• Data Pipeline
• Data Visualization

Generating files

cliUI

filonov \
  --source googleads \
  --googleads.account=ACCOUNT_ID \
  --media-type IMAGE

filonov-ui

Visualizing data

Serveless

Data Visualization is a web application uniformly accessible for all users from http://filonov-ai.web.app.

It's deployed to Firebase Static Hosting and implemented as serverless web application where users can open data files generated with the Data Pipeline.

Self-hosted

Filonov UI can be run locally.

• Install the dependencies

npm install

• Start the app in development mode (hot-code reloading, error reporting, etc.)

  quasar dev
  

• Build the app for production

  quasar build