README.md

Trending Models Dashboard

A Hugging Face Space dashboard for visualizing and analyzing trending model data from the hf-azure-internal/trending-models-analysis dataset.

Overview

This dashboard provides an interactive interface to explore trending machine learning models, their metadata, and their compliance status. Users can filter by date and view models organized by status or trending rank.

Data Source

The dashboard uses data from: hf-azure-internal/trending-models-analysis

Dataset Schema

The models section contains the following columns:

Column	Type	Description
id	string	Model identifier
trending_rank	integer	Current trending position
author	string	Model author/organization
tags	array	Model tags
license	string	Model license type
library_name	string	Framework used (e.g., transformers, diffusers)
gated	boolean	Whether model access is gated
task	string	Model task type
is_in_catalog	boolean	Whether model is in catalog
is_custom_code	boolean	Whether model uses custom code
is_excluded_org	boolean	Whether organization is excluded
is_supported_license	boolean	Whether license is supported
is_supported_library	boolean	Whether library is supported
is_safetensors	boolean	Whether model uses safetensors
is_supported_task	boolean	Whether task is supported
is_securely_scanned	boolean	Whether model is securely scanned
collected_at	datetime	Data collection timestamp
model_status	string	Status: "to add", "blocked", or "added"

Features

User Controls

• Date Filter: Select a specific date to filter models by collected_at timestamp
• Display Mode: Choose between two viewing modes:
  • By Status: Organize models into three sections based on model_status
  • By Trending Rank: Display all models sorted by trending_rank (ascending)

Display Sections

By Status Mode

Models are organized into three sections:

• To Add: Models with model_status = "to add"
• Blocked: Models with model_status = "blocked"
• Added: Models with model_status = "added"

By Trending Rank Mode

• Single section displaying all models sorted by trending_rank

Model Cards

Each model displays:

• Model ID: Raw value from id column
• Status Indicators: Boolean fields shown as emojis
  • 🟢 Green: true values
  • 🔴 Red: false values
  • Fields: is_in_catalog, is_custom_code, is_excluded_org, is_supported_license, is_supported_library, is_safetensors, is_supported_task
• Model Status: Displayed in bold text

Interactive Details

Clicking on any model card reveals additional information:

• License Details: Full license information
• Task Information: Specific task type
• Library Details: Framework/library used
• Author Information: Model author/organization

Technical Requirements

• Will be hosted as a Hugging Face Spaces
• Interactive web interface
• Real-time data filtering
• Smooth animations for detail views
• Responsive design

Development

This project will be implemented as a Hugging Face Space using:

• Python backend for data processing
• Interactive web framework (Gradio/Streamlit)
• Dataset integration via Hugging Face Hub