How santhoshsivang/second-hand-car-shop-customer-review-analysis works
Customer Review AI turns raw review CSVs into satisfaction and journey labels, optionally discovers named topics, and presents the resulting insights in an interactive dashboard with AI-generated keep/improve ideas and editable policy. The dashboard reads processed artifacts, aggregates filters, heatmap cells, and topic bubbles, and uses PostgreSQL and Gemini for idea generation. The README also documents optional AWS deployment. Preprocessing internals and frontend interactions are supported by excerpts; some backend route details and deployment wiring are represented only where the README documents them.
Main entry points: pipeline.py Pipeline orchestration · App.tsx Dashboard shell · api.ts API client · main.py FastAPI routes
Components 25
Offline preprocessing
- Raw reviews
data_preprocessing/raw_data - Pipeline orchestration
data_preprocessing/pipeline.py - Review cleaning
data_preprocessing/cleaning/clean_reviews.py - Satisfaction labels
data_preprocessing/labeling/satisfaction.py - Journey classification
data_preprocessing/journey_classification/classify.py - Output validation
data_preprocessing/validation/validate.py - Processed reviews
data_preprocessing/processed_data - Topic modeling
data_preprocessing/topic_modeling/bertopic_topics.py - Topic catalog
data_preprocessing/topic_modeling/output
Interactive dashboard
- Dashboard shell
app/frontend/src/App.tsx - Filter controls
app/frontend/src/components/FilterSidebar.tsx - Heatmap
app/frontend/src/components/Heatmap.tsx - Topic bubbles
app/frontend/src/components/TopicBubbles.tsx - Idea modal
app/frontend/src/components/IdeaModal.tsx - Policy editor
app/frontend/src/components/PolicyEditor.tsx - API client
app/frontend/src/api.ts
Backend services
- FastAPI routes
app/backend/main.py - Artifact loader
app/backend/data.py - Insight aggregation
app/backend/services.py - Idea generation
app/backend/ideas.py - PostgreSQL
app/backend/db.py
Optional deployment
- AWS deployment
infra/main.tf
Other components
- Analyst
- Gemini
- Browser user
Connections 30
- Analyst → Raw reviews: provides CSV
README.md - Pipeline orchestration → Raw reviews: loads
data_preprocessing/pipeline.py - Pipeline orchestration → Review cleaning: cleans
data_preprocessing/pipeline.py - Pipeline orchestration → Satisfaction labels: labels satisfaction
data_preprocessing/pipeline.py - Pipeline orchestration → Journey classification: classifies stages
data_preprocessing/pipeline.py - Pipeline orchestration → Output validation: validates
data_preprocessing/pipeline.py - Pipeline orchestration → Processed reviews: writes
data_preprocessing/pipeline.py - Pipeline orchestration → Topic modeling: optionally runs
data_preprocessing/pipeline.py - Topic modeling → Gemini: names topics
data_preprocessing/topic_modeling/bertopic_topics.py - Topic modeling → Topic catalog: writes catalog
- Browser user → Dashboard shell: uses
README.md - Dashboard shell → Filter controls: renders
app/frontend/src/App.tsx - Dashboard shell → Heatmap: renders
app/frontend/src/App.tsx - Dashboard shell → Topic bubbles: renders
app/frontend/src/App.tsx - Dashboard shell → Idea modal: opens
README.md - Dashboard shell → Policy editor: renders
app/frontend/src/App.tsx - Filter controls → API client: requests filters
app/frontend/src/components/FilterSidebar.tsx - Heatmap → API client: requests heatmap
app/frontend/src/components/Heatmap.tsx - Topic bubbles → API client: requests topics
app/frontend/src/components/TopicBubbles.tsx - Policy editor → API client: loads and updates
app/frontend/src/components/PolicyEditor.tsx - API client → FastAPI routes: calls API
README.md - FastAPI routes → Artifact loader: loads artifacts
README.md - FastAPI routes → Insight aggregation: serves insights
README.md - FastAPI routes → Idea generation: generates ideas
README.md - Artifact loader → Processed reviews: reads reviews
README.md - Artifact loader → Topic catalog: reads topics
README.md - Idea generation → PostgreSQL: loads policy
README.md - Idea generation → Gemini: requests ideas
README.md - Idea generation → PostgreSQL: saves ideas
README.md - Analyst → AWS deployment: deploys optionally
README.md