Trouver des créateurs avec des audiences qui se chevauchent révèle des opportunités de collaboration et aide les marques à éviter de payer deux influenceurs pour atteindre les mêmes personnes. Ce pipeline compare les listes d'abonnés entre créateurs TikTok, calcule les pourcentages de chevauchement et identifie les segments d'audience communs. Chaque extraction de liste d'abonnés coûte 0,005 $ via les points de terminaison de l'API Scavio TikTok.
Prérequis
- Python 3.8+
- bibliothèque requests
- Une clé API Scavio de scavio.dev
- Noms d'utilisateur des créateurs TikTok à comparer
Parcours
Étape 1: Récupérer les profils des créateurs et des échantillons d'abonnés
Récupérer les données de profil et les listes d'abonnés pour comparaison.
Python
import os, requests, json
from collections import Counter
API_KEY = os.environ['SCAVIO_API_KEY']
TH = {'Authorization': f'Bearer {API_KEY}', 'Content-Type': 'application/json'}
def get_profile(username):
data = requests.post('https://api.scavio.dev/api/v1/tiktok/profile',
headers=TH, json={'username': username}).json()
user = data.get('user', data.get('data', {}).get('user', data))
return {'username': username, 'followers': user.get('followerCount', user.get('fans', 0)),
'following': user.get('followingCount', 0),
'likes': user.get('heartCount', user.get('heart', 0))}
def get_followers(username):
data = requests.post('https://api.scavio.dev/api/v1/tiktok/user/followers',
headers=TH, json={'username': username}).json()
followers = data.get('followers', data.get('data', {}).get('followers', []))
return [f.get('uniqueId', f.get('username', '')) for f in followers if f.get('uniqueId') or f.get('username')]
creators = ['charlidamelio', 'addisonre', 'bellapoarch']
for c in creators:
p = get_profile(c)
print(f'{c}: {p["followers"]:,} followers, {p["likes"]:,} likes')Étape 2: Calculer le chevauchement des abonnés
Comparer les listes d'abonnés entre paires de créateurs.
Python
def overlap(creator_a, creator_b):
followers_a = set(get_followers(creator_a))
followers_b = set(get_followers(creator_b))
if not followers_a or not followers_b:
return {'pair': f'{creator_a} x {creator_b}', 'overlap': 0, 'pct_a': 0, 'pct_b': 0}
shared = followers_a & followers_b
return {
'pair': f'{creator_a} x {creator_b}',
'followers_a': len(followers_a),
'followers_b': len(followers_b),
'shared': len(shared),
'pct_a': len(shared) / len(followers_a) * 100 if followers_a else 0,
'pct_b': len(shared) / len(followers_b) * 100 if followers_b else 0,
'shared_users': list(shared)[:10]
}
result = overlap('charlidamelio', 'addisonre')
print(f'{result["pair"]}: {result["shared"]} shared ({result["pct_a"]:.1f}% of A, {result["pct_b"]:.1f}% of B)')Étape 3: Construire la matrice de chevauchement complète
Comparer toutes les paires de créateurs et construire une matrice.
Python
from itertools import combinations
def overlap_matrix(creators):
pairs = list(combinations(creators, 2))
results = []
print(f'Comparing {len(pairs)} pairs ({len(pairs) * 2} follower pulls)...')
for a, b in pairs:
r = overlap(a, b)
results.append(r)
print(f' {a:20} x {b:20} | shared: {r["shared"]:4} | {r["pct_a"]:.1f}% / {r["pct_b"]:.1f}%')
cost = len(pairs) * 2 * 0.005 # 2 follower pulls per pair
print(f'\nCost: ${cost:.3f} ({len(pairs) * 2} API calls)')
return results
results = overlap_matrix(['charlidamelio', 'addisonre', 'bellapoarch'])
# Highest overlap pair
best = max(results, key=lambda x: x['shared'])
print(f'\nHighest overlap: {best["pair"]} ({best["shared"]} shared followers)')Étape 4: Générer des recommandations de collaboration
Classer les paires de créateurs par chevauchement pour la planification de collaborations de marques.
Python
def recommend(results, strategy='maximize_reach'):
print(f'\n=== Collaboration Recommendations ({strategy}) ===')
if strategy == 'maximize_reach':
# Low overlap = more unique reach
ranked = sorted(results, key=lambda x: x['pct_a'] + x['pct_b'])
print('Pairs with LOWEST overlap (maximum unique reach):')
else:
# High overlap = reinforced messaging
ranked = sorted(results, key=lambda x: x['pct_a'] + x['pct_b'], reverse=True)
print('Pairs with HIGHEST overlap (reinforced messaging):')
for i, r in enumerate(ranked[:5], 1):
avg_overlap = (r['pct_a'] + r['pct_b']) / 2
print(f' {i}. {r["pair"]:45} | overlap: {avg_overlap:.1f}% | shared: {r["shared"]}')
report = {'strategy': strategy, 'recommendations': ranked[:5], 'total_pairs': len(results)}
with open('overlap_report.json', 'w') as f: json.dump(report, f, indent=2, default=str)
print(f'Saved to overlap_report.json')
return report
recommend(results, 'maximize_reach')
recommend(results, 'reinforce_message')Exemple Python
Python
import os, requests
TH = {'Authorization': f'Bearer {os.environ["SCAVIO_API_KEY"]}', 'Content-Type': 'application/json'}
def compare(user_a, user_b):
fa = requests.post('https://api.scavio.dev/api/v1/tiktok/user/followers',
headers=TH, json={'username': user_a}).json()
fb = requests.post('https://api.scavio.dev/api/v1/tiktok/user/followers',
headers=TH, json={'username': user_b}).json()
a = set(f.get('uniqueId', '') for f in fa.get('followers', []))
b = set(f.get('uniqueId', '') for f in fb.get('followers', []))
shared = a & b
print(f'{user_a} x {user_b}: {len(shared)} shared of {len(a)}+{len(b)} ({len(shared)/(len(a) or 1)*100:.1f}%)')
compare('charlidamelio', 'addisonre')Exemple JavaScript
JavaScript
const TH = { 'Authorization': `Bearer ${process.env.SCAVIO_API_KEY}`, 'Content-Type': 'application/json' };
async function compare(userA, userB) {
const [fa, fb] = await Promise.all([
fetch('https://api.scavio.dev/api/v1/tiktok/user/followers', {
method: 'POST', headers: TH, body: JSON.stringify({ username: userA }) }).then(r => r.json()),
fetch('https://api.scavio.dev/api/v1/tiktok/user/followers', {
method: 'POST', headers: TH, body: JSON.stringify({ username: userB }) }).then(r => r.json()),
]);
const a = new Set((fa.followers||[]).map(f => f.uniqueId));
const b = new Set((fb.followers||[]).map(f => f.uniqueId));
const shared = [...a].filter(x => b.has(x));
console.log(`${userA} x ${userB}: ${shared.length} shared of ${a.size}+${b.size}`);
}
compare('charlidamelio', 'addisonre').catch(console.error);Sortie attendue
JSON
charlidamelio: 155,200,000 followers, 11,800,000,000 likes
addisonre: 88,700,000 followers, 5,900,000,000 likes
bellapoarch: 93,400,000 followers, 2,300,000,000 likes
Comparing 3 pairs (6 follower pulls)...
charlidamelio x addisonre | shared: 847 | 42.3% / 38.1%
charlidamelio x bellapoarch | shared: 312 | 15.6% / 14.2%
addisonre x bellapoarch | shared: 523 | 26.2% / 23.8%
Cost: $0.030 (6 API calls)
Highest overlap: charlidamelio x addisonre (847 shared followers)
=== Collaboration Recommendations (maximize_reach) ===
Pairs with LOWEST overlap (maximum unique reach):
1. charlidamelio x bellapoarch | overlap: 14.9% | shared: 312