Final User Interface - View the Live App

YouTube Influencer Discovery

A keyword-driven influencer discovery application built using Streamlit and the YouTube Data API v3.
The tool enables discovery of small to mid-sized YouTube creators across any niche, with pagination, audience-size filtering, and exportable datasets.

Business Problem - As part of the go-to-market (GTM) strategy for an AI-driven test-prep platform, the marketing team needed to build a scalable pool of micro- and mid-tier influencers for content creation, influencer campaigns, video collaborations, and platform reviews. There was no reliable, affordable directory of education-focused creators, and manual discovery was slow, inconsistent, and not repeatable.

Solution- I led the GTM initiative and designed an automated creator-discovery solution to eliminate manual research. Using a rapid, AI-assisted development approach, I built a tool that programmatically identifies YouTube creators based on keyword signals in channel metadata, video titles, and descriptions—allowing the team to surface relevant education-niche creators at scale.

Tech stack used - The solution uses Python as the programming language, Google YouTube data API for Creator and video metadata extraction and Streamlit to host as a light weight web user interface

Results -

Enabled rapid identification of micro- to mid-size education influencers aligned with the GTM strategy
Reduced manual creator discovery effort significantly
Supported outbound outreach workflows by generating structured creator shortlists
Successfully onboarded 5 creators into the influencer program within the initial rollout

🚀 Overview

This application allows users to:

Search YouTube creators by any keyword
Paginate through large result sets
Filter creators by subscriber count
Export structured data for outreach, GTM, or analysis

The project is designed to be API-first, deployment-ready, and easily extensible.

✨ Key Features

Keyword-based creator discovery (channel metadata)
Pagination support (50 results per page)
Subscriber-range filtering
CSV export for downstream workflows
Secure API key handling
Streamlit Cloud deployment support

🧱 Tech Stack

Python
Streamlit
YouTube Data API v3
Pandas

🛠️ Installation

1. Clone the Repository

git clone https://github.com/yourusername/sat-scraper.git
cd sat-scraper

2. Create a Virtual Environment (Optional but Recommended)

python -m venv venv
source venv/bin/activate      # macOS/Linux
venv\Scripts\activate         # Windows

3. Install Dependencies

pip install -r requirements.txt

4. Steps to Generate an API Key

Visit https://console.developers.google.com/
Create a new project
Enable YouTube Data API v3
Go to APIs & Services → Credentials
Create an API key

5. Deployment (Streamlit Cloud)

Push the project to a public GitHub repository
Go to https://streamlit.io/cloud and sign in
Click New App
Select: 4.1. Repository 4.2. Branch (e.g., main) 4.3. Entry file: sat_scraper_app.py
Add the API key under Settings → Secrets
Click Deploy The application will be live at: https://-.streamlit.app

Python Code for YouTube Influencer list extractor

import streamlit as st
from googleapiclient.discovery import build
import pandas as pd
import os
API_KEY = os.getenv("API_KEY")
---- CONFIG ----
#API_KEY = ''   <-- Replace with your actual API key
youtube = build('youtube', 'v3', developerKey=API_KEY)

#---- FUNCTIONS ----
def search_channels(query, max_results=500):
    request = youtube.search().list(
        q=query,
        type='video',
        part='snippet',
        maxResults=max_results
    )
    response = request.execute()
    return response['items']

def get_channel_stats(channel_id):
    request = youtube.channels().list(
        part='statistics,snippet',
        id=channel_id
    )
    response = request.execute()
    if response['items']:
        stats = response['items'][0]
        return {
            'Channel Name': stats['snippet']['title'],
            'Channel ID': channel_id,
            'Subscribers': int(stats['statistics'].get('subscriberCount', 0)),
            'Video Count': stats['statistics'].get('videoCount', 'N/A'),
            'Description': stats['snippet'].get('description', ''),
            'URL': f"https://www.youtube.com/channel/{channel_id}"
        }
    return {}

 #---- STREAMLIT UI ----
st.title("🎓 YouTube Influencer Finder")

search_query = st.text_input("Search Keyword", value="SAT prep")
max_subs = st.slider("Max Subscriber Count", min_value=100, max_value=500000, value=250000, step=1000)
run_search = st.button("Search YouTube Influencers")

if run_search:
    with st.spinner("Searching..."):
        results = search_channels(search_query)
        influencer_data = []

        for item in results:
            channel_id = item['snippet']['channelId']
            stats = get_channel_stats(channel_id)
            if 100 <= stats.get('Subscribers', 0) <= max_subs:
                influencer_data.append(stats)

        if influencer_data:
            df = pd.DataFrame(influencer_data)
            st.success(f"Found {len(df)} influencers.")
            st.dataframe(df)
            csv = df.to_csv(index=False).encode('utf-8')
            st.download_button("Download CSV", csv, "youtube_influencers.csv", "text/csv")
        else:
            st.warning("No influencers found in that range.")

Keyword-Based Influencer Discovery