Simon-Pierre Boucher
All apps & projects
Web Platform
JavaScript

InfluenceursQC

The data-driven ranking of Quebec's influencers and content creators.

InfluenceursQC ranks 376 Quebec influencers across six platforms, with sourced editorial dossiers and transparent net-worth estimates in CAD. Live at www.influenceursqc.com.

376

Influencers ranked by cumulative audience

6

Platforms tracked: Instagram, TikTok, YouTube, Facebook, X, Twitch

15

Content categories

3,732

Gallery images, each linked to its source

1,464

Sourced facts (888 of them dated)

1,053

News items citing their sources

Overview

InfluenceursQC is a web platform and complete dataset of Quebec's notable influencers and content creators: 376 profiles with multi-platform metrics across Instagram, TikTok, YouTube, Facebook, X, and Twitch. Beyond the global ranking by cumulative audience, it offers per-platform rankings, 15 content categories, and rich individual profile pages combining global and category ranks, an audience bar chart per platform, a masonry photo gallery, a dated career timeline, sourced highlights, a news feed, and an estimated net worth range in Canadian dollars.

The dataset is built by a documented pipeline: roughly 28 French and English Firecrawl web searches over general and niche rankings, about 32 scraped source pages (Feedspot, Collabstr, Favikon, Modash, Quebec agencies), and structured extraction by parallel agents into normalized JSON per source. A Python merge script deduplicates on normalized names, aliases, and handles, keeps the maximum value per metric, filters out non-Quebec profiles, and assigns audience tiers. Enrichment adds 376 image searches yielding 3,732 gallery images and 376 web-and-news searches producing 1,464 sourced facts and 1,053 news items — each one citing its source.

The application itself is deliberately simple: a zero-build Express server-side-rendered app whose server.js holds the server, templates, net-worth computation, and full-text search over the embedded JSON dataset, exposed through an advanced search page and a JSON API at /api/influenceurs. The transparent net-worth model publishes its per-platform CAD rates on a public methodology page. In production it runs on MacLustr cluster node M3U96a under PM2, with an ngrok tunnel serving www.influenceursqc.com; data refreshes flow through merge.py, rsync, and a PM2 restart.

Key Features

Global and platform rankings

376 influencers ranked by cumulative audience with stat tiles and magnitude bars, plus dedicated tops for Instagram, TikTok, YouTube, Facebook, X, and Twitch.

Rich profile pages

Each profile shows its global and category ranks, a per-platform audience bar chart, direct links to every account, a dated career timeline, highlights, and a news feed — every fact cites its source.

15 category rankings

Comedy, food, fashion, gaming, dance, and more — each of the 15 domains gets its own ranking with cumulative audience totals.

Advanced search

Full-text search that includes the harvested web content (facts and news), with platform, category, tier, and minimum-audience filters, sortable by audience, net worth, or name.

Transparent net-worth estimates

A CAD range per profile computed from published per-platform rates (Instagram $0.20, TikTok $0.08, YouTube $0.35, Facebook $0.05, X $0.03, Twitch $0.50 per follower per year), tier and category multipliers, and active years — fully documented on /methodologie.

Open JSON API

The complete dataset is served at /api/influenceurs alongside a /health endpoint, with CSV and JSON exports maintained in the repository.

How It Works

  1. Broad web research

    About 28 French and English Firecrawl searches sweep general and niche influencer rankings, and roughly 32 source pages (Feedspot, Collabstr, Favikon, Modash, Quebec agencies) are scraped as raw material.

  2. Structured extraction

    Parallel agents turn each group of sources into normalized JSON records, stored as intermediate extractions before merging.

  3. Merge and deduplication

    scripts/merge.py keys profiles on normalized name, aliases, and handles, keeps the maximum value per metric, filters out non-Quebec profiles, and assigns audience tiers to produce the final deduplicated dataset.

  4. Enrichment

    Targeted searches fill missing platforms for major profiles, 376 image searches build the 3,732-image gallery, and 376 web-and-news searches produce the sourced facts and news items in the editorial dossiers.

  5. Serve and deploy

    A zero-build Express SSR server renders rankings, profiles, search, and the net-worth model from embedded JSON, running on MacLustr node M3U96a under PM2 behind an ngrok tunnel at www.influenceursqc.com.

Tech Stack

Web App

Node.js + Express (SSR, zero build)
Embedded JSON dataset
Full-text search at /recherche
JSON API at /api/influenceurs

Data Pipeline

Firecrawl Search & Scrape
Parallel extraction agents
Python merge/dedup scripts
CSV + JSON dataset exports

Deployment

MacLustr node M3U96a
PM2 (app + tunnel processes)
ngrok tunnel
www.influenceursqc.com

Highlights

  • 376 Quebec influencers profiled with multi-platform metrics across Instagram, TikTok, YouTube, Facebook, X, and Twitch.
  • Every editorial item cites its source: 1,464 sourced facts (888 dated) and 1,053 news items compiled per profile.
  • 3,732 gallery images collected through Firecrawl image search, displayed in masonry layouts and linked back to their sources.
  • A transparent net-worth model with published per-platform CAD rates and tier multipliers, documented on a public /methodologie page.
  • Advanced full-text search covers the harvested web content itself, with platform, category, tier, and minimum-audience filters.
  • The complete dataset is open through a JSON API at /api/influenceurs, with the whole app running as a single zero-build Express SSR server.

Explore InfluenceursQC

Ranking Quebec's influencers across six platforms — the full source is on GitHub.