Customer Feedback Intelligence Dashboard

Triage customer feedback with a saved sentiment model, surface the most urgent reviews, and compare current behavior against benchmark and transfer results.

Active Inference Model

Backend: scikit-learn

Artifact: artifacts/models/tfidf_logreg_imdb.joblib

IMDb Benchmark

Test accuracy: 0.9015

Macro F1: 0.9015

Amazon Transfer

Accuracy: 0.8565

Macro F1: 0.8560

Paste multiple reviews separated by blank lines. The dashboard scores sentiment, estimates triage priority, and groups the batch into themes. CSV is the most common spreadsheet export format. Uploads also accept TSV, JSON, JSONL/NDJSON, TXT, or Markdown files.
Predicted Label Filter
Priority Filter
Sort Reviews By

Filters apply to the most recent analysis result without rerunning the model.

Theme Clusters

Theme Clusters

Priority Queue

Priority Queue