Niteni Social Media Analytics Documentation

Niteni adalah platform analisis ulasan dan media sosial berbasis Artificial Intelligence yang mengubah review pelanggan menjadi insight strategis untuk manajemen.

Google Review
↓
Repository Builder
↓
Emotion Analysis
↓
Meaning Extraction
↓
Thematic Clustering
↓
Cluster Emotion
↓
Cluster Naming
↓
SWOT Analysis
↓
Business Model Canvas
↓
Balanced Scorecard
↓
Competitor Analysis

1. Build Repository

Tahap pertama mengubah review mentah menjadi repository analitik.

Input

reviews.json

Proses

Split review menjadi kalimat.
Filter panjang kalimat.
Emotion Analysis menggunakan NRC Dictionary.
Sentiment Analysis.
SWOT Dictionary Tagging.
Penyimpanan repository.

Output

repository.json
repository_metadata.json

Node Script

node /opt/scrapgmap/tools/build_repository.js

Dashboard

Build Repository

2. Emotion Analysis

Analisis emosi dilakukan pada level kalimat menggunakan NRC Emotion Lexicon.

Emosi yang Didukung

Positive Emotion	Negative Emotion
Joy	Anger
Trust	Fear
Anticipation	Sadness
Surprise	Disgust

Output

positive
negative
joy
trust
anticipation
surprise
anger
fear
sadness
disgust

3. Build Meaning

Mengubah kalimat review menjadi makna singkat menggunakan AI Summarization.

Contoh

Sentence	Meaning
Harga tiket masuk terlalu mahal untuk fasilitas yang diberikan	harga tiket mahal
Staf sangat ramah dan membantu selama menginap	staf ramah

API

https://translate.niteni.web.id/summarize/batch

Output

meanings.json
meanings_metadata.json

Node Script

node /opt/scrapgmap/tools/build_meanings.js

Dashboard

Build Meaning

4. Build Cluster

Mengelompokkan makna yang memiliki kemiripan semantik menjadi tema yang sama.

Engine

Sentence Embedding
+
HDBSCAN GPU (cuML)

Contoh

harga tiket mahal
biaya masuk mahal
tiket terlalu mahal
harga masuk mahal

Menjadi:

Cluster #14
Harga Mahal

API

https://cluster.niteni.web.id/api/thematic/initial-gpu

Output

cluster_mapping.json
clusters.json
cluster_samples.json
clusters_metadata.json

Node Script

node /opt/scrapgmap/tools/build_clusters.js

Per Place

node /opt/scrapgmap/tools/build_clusters.js PLACE_ID

Dashboard

Build Cluster

5. Cluster Emotion Analysis

Menghitung emosi pada setiap cluster untuk mengetahui penyebab utama kepuasan atau ketidakpuasan pelanggan.

Contoh

Cluster:
Staff Service

Mention:
119

Positive:
110

Negative:
9

Tujuan

Menentukan Strength.
Menentukan Weakness.
Membantu SWOT Analysis.
Membantu Competitor Analysis.

6. Cluster Naming (LLM)

Memberikan nama bisnis yang mudah dipahami untuk setiap cluster.

Sebelum	Sesudah
Cluster 77	Staff Service
Cluster 14	Price
Cluster 44	Hotel Experience

7. SWOT Analysis

Komponen	Sumber Data
Strength	Cluster positif dominan
Weakness	Cluster negatif dominan
Opportunity	Cluster netral yang sering muncul
Threat	Cluster negatif yang meningkat dari waktu ke waktu

Contoh

Strength
- Staff Service
- Breakfast
- Room Quality

Weakness
- Price
- Parking
- Beach Access

Opportunity
- Kids Club
- Meeting Room

Threat
- Competitor Pricing
- Beach Cleanliness

8. Business Model Canvas Analysis

BMC Component	Sumber Cluster
Value Proposition	Staff, Food, View, Service
Customer Relationship	Response Time, Follow Up
Channels	Booking, Website, WhatsApp
Customer Segment	Family, Couple, Business
Revenue Streams	Room, Food, Spa
Key Resources	Staff, Building, Beach
Key Activities	Housekeeping, Restaurant
Key Partners	Travel Agent, Vendor
Cost Structure	Price, Parking, Additional Cost

9. Balanced Scorecard Analysis

Perspective	Cluster Example
Financial	Price, Promotion, Value
Customer	Service, Room, Food
Internal Process	Check-In, Cleaning, Reservation
Learning & Growth	Staff Competency, Training

10. Competitor Analysis

Membandingkan beberapa Place ID yang berada pada industri yang sama.

Contoh

Hotel A
Hotel B
Hotel C

Dibandingkan berdasarkan:

Strength.
Weakness.
Food.
Room.
Service.
Price.
Location.

Repository Structure

repository/
└── PLACE_ID/
    ├── reviews.json
    ├── repository.json
    ├── meanings.json
    ├── meanings_metadata.json
    ├── cluster_mapping.json
    ├── clusters.json
    ├── cluster_samples.json
    └── clusters_metadata.json

Roadmap

Feature	Status
Repository Builder	✅ Completed
Emotion Analysis	✅ Completed
Meaning Builder	✅ Completed
Cluster Builder	✅ Completed
Cluster Emotion	✅ Completed
Cluster Naming	✅ Completed
SWOT Analysis	✅ Completed
Business Model Canvas	✅ Completed
Balanced Scorecard	🔄 Planned
Competitor Analysis	🔄 Planned