📊 Customer Feedback Sentiment Dashboard

An AI-powered dashboard to analyze customer feedback sentiment, group themes, and generate reports. Built with NestJS for scalability and Mistral AI for NLP.

🛠️ Stack

Core

Backend: NestJS (TypeScript)
Database: PostgreSQL (Drizzle ORM)
Auth: JWT + Redis (sessions/rate-limiting)
AI: Mistral API (Sentiment Analysis + Topic Clustering)

Tools

Validation: Zod + NestJS-Zod
API Docs: Swagger
Testing: Jest + Supertest
Linting/Formatting: Biome
File Processing: Multer (uploads), PapaParse (CSV)
Monitoring: Prometheus + Grafana + Loki
Deployment: Docker + Docker Compose
CI/CD: GitHub Actions
PDF Generation: PDF-lib
CSV Export: JSON2CSV
Real-time: Socket.IO

🛠 How to Run

1. Clone the repository

git clone https://github.com/Xayrulloh/Feedback-Sentiment.git && cd Feedback-Sentiment

2. Create .env file:

DATABASE_URL='postgresql://<user>:<password>@localhost:5432/<database>'
PORT=XXXX

# JWT
JWT_SECRET=

# AI
OPENAI_API_KEY

# POSTGRES
POSTGRES_USER
POSTGRES_PASSWORD
POSTGRES_DB
POSTGRES_PORT

# Redis
REDIS_HOST
REDIS_PORT

3. Run Database

docker compose up -d  # Starts PostgreSQL + Redis
npx drizzle-kit push:pg  # Apply DB schema

4. Start Server

pnpm run start:dev

5. Swagger Docs and server address

http://localhost:${PORT}/${GLOBAL_PREFIX} server address
http://localhost:${PORT}/${GLOBAL_PREFIX}/docs swagger docs

🧠 How It Works

1. Feedback Ingestion:

CSV/text feedbacks input → AI Processing → Stored in PostgreSQL.

2. AI Processing:

Mistral API tags sentiment → Groups similar feedback via NLP -> Tells confidence

3. Dashboard:

Filter by sentiment, group themes, download reports, bar charts.

4. Admin Dashboard:

Get users -> Disable users -> Suspend users ->
Get suspicious activity (Registration/Upload/Download/API)
Rate limit (API/Upload/Download/Login)

🌐 Endpoints

🔐 Authentication Endpoints

1. User Registration

Path: POST /auth/(register|login)
Flow Register/Login:

sequenceDiagram
    Frontend->>+Backend: POST /auth/register {email, password}
    Backend->>+UserStatusGuard: Check suspended/disabled
    Backend->>+RateLimitGuard: Check user rate limits and increment (create activity in suspicious table)
    RateLimitGuard->>+Websocket: Send admins if user hits limits
    Backend->>+MetricsMiddleware: (Api Usage, Error Rates, Uploads)
    Backend->>+Websocket: Active users
    Backend->>+Backend: Zod Validation
    Backend->>+DB: Checking email and password
    alt Not valid
        DB-->>-Backend: Error (Incorrect password or Email exist)
        Backend-->>-Frontend: 409 | 401
    else Valid
        Backend->>DB: Hash password, save user (role=user)
        Backend->>Frontend: 201 + JWT
    end

2. Admin Registration

Path: POST /auth/(register|login)/admin
Flow Register/Login:

sequenceDiagram
    Frontend->>+Backend: POST /auth/register {email, password}
    Backend->>+MetricsMiddleware: (Api Usage, Error Rates, Uploads)
    Backend->>+Websocket: Active users
    Backend->>+Backend: Zod Validation
    Backend->>+DB: Checking email and password
    alt Not valid
        DB-->>-Backend: Error (Incorrect password or Email exist)
        Backend-->>-Frontend: 409 | 401
    else Valid
        Backend->>DB: Hash password, save user (role=user)
        Backend->>Frontend: 201 + JWT
    end

📥 Feedback Endpoints

1. Feedback Manual

Path: POST /feedback/manual Flow:

sequenceDiagram
    Frontend->>+Backend: POST /feedback/manual (array of feedbacks)
    Backend->>+UserStatusGuard: Check suspended/disabled
    Backend->>+RateLimitGuard: Check user rate limits and increment (create activity in suspicious table)
    RateLimitGuard->>+Websocket: Send admins if user hits limits
    Backend->>+MetricsMiddleware: (Api Usage, Error Rates, Uploads)
    Backend->>Backend: Zod Validation
    Backend->>DB: Check whether it's already exist or not
    alt Exist
        DB-->>-Backend: Get that feedbacks
        Backend-->>-Frontend: 201
    else Create
        Backend->>+AIService: Analyze (confidence/sentiment/group)
        AIService->>+DB: Create new one and return
        DB-->>-AIService: Success
        AIService-->>-Backend: Processed
        Backend-->>-Frontend: 201
    end

2. Upload Feedback via CSV

Path: POST /feedback/upload
Flow:

sequenceDiagram
    Frontend->>+Backend: POST /feedback/upload (file which has array of feedbacks)
    Backend->>+UserStatusGuard: Check if user suspended/disabled
    Backend->>+RateLimitGuard: Check user rate limits and increment (create activity in suspicious table)
    RateLimitGuard->>+Websocket: Send admins if user hits limits
    Backend->>+MetricsMiddleware: (Api Usage, Error Rates, Uploads)
    Backend->>Backend: Parsing and Zod Validation
    Backend->>+FeedbackManual: Same happens with manual with the rest
    Backend-->>-Frontend: 201 Created with results

3. Get Feedback by ID

Path: GET /feedback/:id
Flow:

sequenceDiagram
    Frontend->>+Backend: GET /feedback/:id
    Backend->>+UserStatusGuard: Check if user suspended/disabled
    Backend->>+RateLimitGuard: Check user rate limits and increment (create activity in suspicious table)
    RateLimitGuard->>+Websocket: Send admins if user hits limits
    Backend->>+MetricsMiddleware: (Api Usage, Error Rates, Uploads)
    Backend->>Backend: Zod Validation
    Backend->>+DB: Get feedback by id
    DB-->>-Backend: Feedback
    Backend-->>-Frontend: 200

4. Filter Feedback by Sentiment

Path: GET /feedback
Flow:

sequenceDiagram
    Frontend->>+Backend: GET /feedback?sentiment=positive,negative
    Backend->>+UserStatusGuard: Check if user suspended/disabled
    Backend->>+RateLimitGuard: Check user rate limits and increment (create activity in suspicious table)
    RateLimitGuard->>+Websocket: Send admins if user hits limits
    Backend->>+MetricsMiddleware: (Api Usage, Error Rates, Uploads)
    Backend->>Backend: Zod Validation
    Backend->>+DB: Query with Sentiment Filters
    DB-->>-Backend: Filtered Results
    Backend-->>-Frontend: 200 (Paginated Data)

5. Get Feedback Summary

Path: GET /feedback/sentiment-summary
Flow:

sequenceDiagram
    Frontend->>+Backend: GET /feedback/sentiment-summary
    Backend->>+UserStatusGuard: Check if user suspended/disabled
    Backend->>+RateLimitGuard: Check user rate limits and increment (create activity in suspicious table)
    RateLimitGuard->>+Websocket: Send admins if user hits limits
    Backend->>+MetricsMiddleware: (Api Usage, Error Rates, Uploads)
    Backend->>Backend: Zod Validation
    Backend->>+DB: Query with userId
    DB-->>-Backend: Grouped Results
    Backend-->>-Frontend: 200

6. Get Feedback Grouped

Path: GET /feedback/grouped Flow:

sequenceDiagram
    Frontend->>+Backend: GET /feedback/grouped
    Backend->>+UserStatusGuard: Check if user suspended/disabled
    Backend->>+RateLimitGuard: Check user rate limits and increment (create activity in suspicious table)
    RateLimitGuard->>+Websocket: Send admins if user hits limits
    Backend->>+MetricsMiddleware: (Api Usage, Error Rates, Uploads)
    Backend->>Backend: Zod Validation
    Backend->>+DB: Query with userId
    DB-->>-Backend: Grouped by summary 
    Backend-->>-Frontend: 200

7. Get Feedback Report file (CSV/PDF)

Path: GET /feedback/report Flow:

sequenceDiagram
    Frontend->>+Backend: GET /feedback/report
    Backend->>+UserStatusGuard: Check if user suspended/disabled
    Backend->>+RateLimitGuard: Check user rate limits and increment (create activity in suspicious table)
    RateLimitGuard->>+Websocket: Send admins if user hits limits
    Backend->>+MetricsMiddleware: (Api Usage, Error Rates, Uploads)
    Backend->>Backend: Zod Validation
    Backend->>+DB: Fetch all feedback
    DB-->>-Backend: Raw data
    alt PDF Format
        Backend->>+PDF-lib: Generate PDF
        PDF-lib-->>-Backend: PDF buffer
    else CSV Format
        Backend->>+json2csv: Convert to CSV
        json2csv-->>-Backend: CSV string
    end
    Backend-->>-Frontend: File download

📁 File Endpoints

1. Get Files

Path: GET /files
Flow:

sequenceDiagram
    Frontend->>+Backend: GET /files
    Backend->>+UserStatusGuard: Check if user suspended/disabled
    Backend->>+RateLimitGuard: Check user rate limits and increment (create activity in suspicious table)
    RateLimitGuard->>+Websocket: Send admins if user hits limits
    Backend->>+MetricsMiddleware: (Api Usage, Error Rates, Uploads)
    Backend->>Backend: Zod Validation
    Backend->>+DB: Fetch all user files (Paginated)
    DB-->>-Backend: Files
    Backend-->>-Frontend: 200

2. Delete File

Path: DELETE /files/:id
Flow:

sequenceDiagram
    Frontend->>+Backend: DELETE /files/:id
    Backend->>+UserStatusGuard: Check if user suspended/disabled
    Backend->>+RateLimitGuard: Check user rate limits and increment (create activity in suspicious table)
    RateLimitGuard->>+Websocket: Send admins if user hits limits
    Backend->>+MetricsMiddleware: (Api Usage, Error Rates, Uploads)
    Backend->>Backend: Zod Validation
    Backend->>+DB: Delete file
    DB-->>-Backend: Deleted
    Backend-->>-Frontend: 200

🤨 User Endpoints

1. Get Users

Path: GET /users
Flow:

sequenceDiagram
    Frontend->>+Backend: GET /users
    Backend->>+UserStatusGuard: Check if user suspended/disabled
    Backend->>+RateLimitGuard: Check user rate limits and increment (create activity in suspicious table)
    RateLimitGuard->>+Websocket: Send admins if user hits limits
    Backend->>+MetricsMiddleware: (Api Usage, Error Rates, Uploads)
    Backend->>Backend: Zod Validation
    Backend->>+DB: Fetch all users (Paginated)
    DB-->>-Backend: Users
    Backend-->>-Frontend: 200

2. Search Users

Path: GET /users/search
Flow:

sequenceDiagram
    Frontend->>+Backend: GET /users/search?email=example
    Backend->>+UserStatusGuard: Check if user suspended/disabled
    Backend->>+RateLimitGuard: Check user rate limits and increment (create activity in suspicious table)
    RateLimitGuard->>+Websocket: Send admins if user hits limits
    Backend->>+MetricsMiddleware: (Api Usage, Error Rates, Uploads)
    Backend->>Backend: Zod Validation
    Backend->>+DB: Search users by email
    DB-->>-Backend: Users (limit 5)
    Backend-->>-Frontend: 200

🔒 Admin Endpoints

1. Disable User

Path: POST /admin/disable/:userId Flow:

sequenceDiagram
    Frontend->>+Backend: POST /admin/disable/:userId
    Backend->>+Backend: Check the user is admin
    Backend->>+DB: Disable user
    DB-->>-Backend: Disabled
    Backend-->>-Frontend: 200

2. Suspend User

Path: POST /admin/suspend/:userId Flow:

sequenceDiagram
    Frontend->>+Backend: POST /admin/suspend/:userId
    Backend->>+Backend: Check the user is admin
    Backend->>+DB: Suspend user | Enable user
    DB-->>-Backend: Suspended | Enabled 
    Backend-->>-Frontend: 200

3. Metrics

Path: GET /admin/metrics Flow:

sequenceDiagram
    Frontend->>+Backend: GET /admin/metrics
    Backend->>+Backend: Check the user is admin
    Backend->>+DB: Take metrics (uploads, api usage, error rates)
    DB-->>-Backend: Metrics
    Backend-->>-Frontend: 200

4. Rate Limit

Path: GET /admin/rate-limit Flow:

sequenceDiagram
    Frontend->>+Backend: GET /admin/rate-limit
    Backend->>+Backend: Check the user is admin
    Backend->>+DB: Take rate limits (api, upload, download, login)
    DB-->>-Backend: RateLimits
    Backend-->>-Frontend: 200

5. Rate Limit

Path: PATCH /admin/rate-limit Flow:

sequenceDiagram
    Frontend->>+Backend: PATCH /admin/rate-limit
    Backend->>+Backend: Check the user is admin
    Backend->>+DB: Path rate limits (api, upload, download, login)
    DB-->>-Backend: RateLimits
    Backend-->>-Frontend: 200

6. Suspicious Activity

Path: GET /admin/suspicious-activities Flow:

sequenceDiagram
    Frontend->>+Backend: GET /admin/suspicious-activities
    Backend->>+Backend: Check the user is admin
    Backend->>+DB: Take suspicious activities
    DB-->>-Backend: SuspiciousActivities
    Backend-->>-Frontend: 200

📜 Sample Feedback Endpoints

1 Sample. Feedback Manual

Path: POST /sample/feedback/manual Flow:

sequenceDiagram
    Frontend->>+Backend: POST /sample/feedback-manual (array of feedbacks)
    Backend->>+UserStatusGuard: Check suspended/disabled
    Backend->>+RateLimitGuard: Check user rate limits and increment (create activity in suspicious table)
    RateLimitGuard->>+Websocket: Send admins if user hits limits
    Backend->>+MetricsMiddleware: (Api Usage, Error Rates, Uploads)
    Backend->>Backend: Zod Validation
    Backend->>MockData: Mock feedbacks
    Backend-->>-Frontend: 201

2. Sample Get Feedback by ID

Path: GET /sample/feedback/:id
Flow:

sequenceDiagram
    Frontend->>+Backend: GET /sample/feedback-:id
    Backend->>+UserStatusGuard: Check if user suspended/disabled
    Backend->>+RateLimitGuard: Check user rate limits and increment (create activity in suspicious table)
    RateLimitGuard->>+Websocket: Send admins if user hits limits
    Backend->>+MetricsMiddleware: (Api Usage, Error Rates, Uploads)
    Backend->>Backend: Zod Validation
    Backend->>+MockData: Mock Feedback
    Backend-->>-Frontend: 200

4. Sample Filter Feedback by Sentiment

Path: GET /feedback
Flow:

sequenceDiagram
    Frontend->>+Backend: GET /feedback?sentiment=positive,negative
    Backend->>+UserStatusGuard: Check if user suspended/disabled
    Backend->>+RateLimitGuard: Check user rate limits and increment (create activity in suspicious table)
    RateLimitGuard->>+Websocket: Send admins if user hits limits
    Backend->>+MetricsMiddleware: (Api Usage, Error Rates, Uploads)
    Backend->>Backend: Zod Validation
    Backend->>+MockData: Mock Feedback
    Backend-->>-Frontend: 200 (Paginated Data)

5. Sample Get Feedback Summary

Path: GET /sample/feedback-sentiment-summary
Flow:

sequenceDiagram
    Frontend->>+Backend: GET /sample/feedback-sentiment-summary
    Backend->>+UserStatusGuard: Check if user suspended/disabled
    Backend->>+RateLimitGuard: Check user rate limits and increment (create activity in suspicious table)
    RateLimitGuard->>+Websocket: Send admins if user hits limits
    Backend->>+MetricsMiddleware: (Api Usage, Error Rates, Uploads)
    Backend->>Backend: Zod Validation
    Backend->>+MockData: Mock Feedback
    Backend-->>-Frontend: 200

6. Sample Get Feedback Grouped

Path: GET /sample/feedback-grouped Flow:

sequenceDiagram
    Frontend->>+Backend: GET /sample/feedback-grouped
    Backend->>+UserStatusGuard: Check if user suspended/disabled
    Backend->>+RateLimitGuard: Check user rate limits and increment (create activity in suspicious table)
    RateLimitGuard->>+Websocket: Send admins if user hits limits
    Backend->>+MetricsMiddleware: (Api Usage, Error Rates, Uploads)
    Backend->>Backend: Zod Validation
    Backend->>+MockData: Mock Feedback
    Backend-->>-Frontend: 200

7. Sample Get Feedback Report file (CSV/PDF)

Path: GET /sample/feedback-report Flow:

sequenceDiagram
    Frontend->>+Backend: GET /sample/feedback-report
    Backend->>+UserStatusGuard: Check if user suspended/disabled
    Backend->>+RateLimitGuard: Check user rate limits and increment (create activity in suspicious table)
    RateLimitGuard->>+Websocket: Send admins if user hits limits
    Backend->>+MetricsMiddleware: (Api Usage, Error Rates, Uploads)
    Backend->>Backend: Zod Validation
    Backend->>+MockData: Mock Feedback
    Backend-->>-Frontend: File download

🗃️ Database Schema

erDiagram
    users ||--o{ users_feedbacks : "1:N"
    users ||--o{ files : "1:N"
    users ||--o{ suspicious_activity : "1:N"
    feedbacks ||--o{ users_feedbacks : "1:N"
    files ||--o{ users_feedbacks : "1:N"

    users {
        string id PK "uuid"
        string email "unique"
        string password_hash
        enum role "user|admin"
        boolean is_suspended
        timestamp created_at
        timestamp updated_at
        timestamp deleted_at "nullable"
    }

    feedbacks {
        string id PK "uuid"
        string content_hash "unique, sha256"
        text content
        enum sentiment "positive|neutral|negative|unknown"
        integer confidence "0-100"
        text summary "AI-generated"
        timestamp created_at
        timestamp updated_at
        timestamp deleted_at "nullable"
    }

    users_feedbacks {
        string user_id FK,PK
        string feedback_id FK,PK
        string file_id FK "nullable"
        timestamp created_at
    }

    files {
        string id PK "uuid"
        string user_id FK
        string name
        string mime_type
        bigint size "bytes"
        integer row_count "nullable"
        string extension
        timestamp created_at
        timestamp updated_at
        timestamp deleted_at "nullable"
    }

    rate_limits {
        string id PK "uuid"
        enum target "api|upload|download|login"
        integer limit
        timestamp created_at
        timestamp updated_at
        timestamp deleted_at "nullable"
    }

    suspicious_activity {
        string id PK "uuid"
        string user_id FK "nullable"
        string email "nullable"
        string ip "nullable"
        enum action "api|upload|download|login"
        enum error "too_many_*"
        text details "nullable"
        timestamp created_at
        timestamp updated_at
        timestamp deleted_at "nullable"
    }

Name		Name	Last commit message	Last commit date
Latest commit History 384 Commits
.github/workflows		.github/workflows
.vscode		.vscode
src		src
test		test
.dockerignore		.dockerignore
.env.example		.env.example
.gitignore		.gitignore
.prettierrc		.prettierrc
Dockerfile		Dockerfile
LICENSE		LICENSE
README.md		README.md
biome.json		biome.json
docker-compose.yml		docker-compose.yml
drizzle.config.ts		drizzle.config.ts
entrypoint.sh		entrypoint.sh
eslint.config.mjs		eslint.config.mjs
nest-cli.json		nest-cli.json
package.json		package.json
pnpm-lock.yaml		pnpm-lock.yaml
prometheus.yml		prometheus.yml
promtail-config.yml		promtail-config.yml
tsconfig.build.json		tsconfig.build.json
tsconfig.json		tsconfig.json

Folders and files

Latest commit

History

Repository files navigation

📊 Customer Feedback Sentiment Dashboard

🛠️ Stack

Core

Tools

🛠 How to Run

1. Clone the repository

2. Create .env file:

3. Run Database

4. Start Server

5. Swagger Docs and server address

🧠 How It Works

1. Feedback Ingestion:

2. AI Processing:

3. Dashboard:

4. Admin Dashboard:

🌐 Endpoints

🔐 Authentication Endpoints

1. User Registration

2. Admin Registration

📥 Feedback Endpoints

1. Feedback Manual

2. Upload Feedback via CSV

3. Get Feedback by ID

4. Filter Feedback by Sentiment

5. Get Feedback Summary

6. Get Feedback Grouped

7. Get Feedback Report file (CSV/PDF)

📁 File Endpoints

1. Get Files

2. Delete File

🤨 User Endpoints

1. Get Users

2. Search Users

🔒 Admin Endpoints

1. Disable User

2. Suspend User

3. Metrics

4. Rate Limit

5. Rate Limit

6. Suspicious Activity

📜 Sample Feedback Endpoints

1 Sample. Feedback Manual

2. Sample Get Feedback by ID

4. Sample Filter Feedback by Sentiment

5. Sample Get Feedback Summary

6. Sample Get Feedback Grouped

7. Sample Get Feedback Report file (CSV/PDF)

🗃️ Database Schema

About

Topics

Resources

License

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Packages