csg
Résumé ↗
Menu
← All projects

NLP · Product engineering / Hosted demo

Review Tag Generator

Product feedback, down to the evidence.

Two fine-tuned models turn reviews into feature-level sentiment. Every insight keeps its source text, with separate reviewer corrections and annotations.

DistilBERT · Python · React · SQLite

View repository ↗Open hosted demo ↗

01 / THE PROBLEM

What needed solving.

A star rating hides which part of a product worked or failed. Feature-level analysis needs both an aspect span and its sentiment, while keeping the original evidence inspectable. A useful product also needs to distinguish model predictions from human corrections.

02 / FOLLOW THE WORKFLOW

See how it fits together.

A simplified example of the documented flow. Sample outputs are authored illustrations, not live model predictions.

Review Tag GeneratorIllustrative workflow

Synthetic review

“The battery lasts all day.”

Start with a product review.

Aspect span: battery

The [battery] lasts all day.

Identify the feature mentioned in the text.

Illustrative label, not live inference

Battery → Positive

Attach sentiment to the extracted aspect.

Preview first → explicitly save

Battery life · “lasts all day”

Keep the supporting text beside the feature.

03 / ENGINEERING

The decisions behind it.

Split extraction from sentiment

One fine-tuned DistilBERT checkpoint extracts aspect spans; another classifies sentiment. Deterministic rules normalize known aliases, generate tags and aggregate saved predictions into product insights.

Preview before changing totals

A review can be analyzed without immediately contributing to saved product counts. Explicit saving, durable bulk-import jobs and row-error exports make data changes visible.

Preserve human decisions separately

Accept, edit or reject a suggestion, correct sentiment, or annotate missed spans. Original predictions and decision history remain traceable. These edits do not automatically retrain the model or change product totals.

Freeze evaluation choices

The full-review sentiment policy was selected on validation before the held-out test. Reported extraction and sentiment scores have different denominators and do not establish complete tag accuracy.

04 / EVIDENCE & LIMITS

Claims with context.

70.62%Exact-span F1 · 1,616 test reviews
74.96%Sentiment Macro F1 · 1,445 gold pairs

The saved checkpoints reproduce the documented frozen-test results: extraction precision 64.85%, recall 77.51%, and sentiment accuracy 82.49%. Sentiment evaluation uses gold aspects. A separate ten-request warm local smoke measurement reported 566.45 ms median preview time on Apple M4, including scheduling and polling; it is not a production latency benchmark.

Current limits & next validation

End-to-end canonical-tag quality is not yet established. Full-review context can confuse contrasting aspect sentiments. The hosted backend may need time to wake; optional cloud AI explanations are disabled according to the README. The retailer adapter is experimental.

05 / READ THE SOURCE

Go deeper.

Based on public repository documentation reviewed on 7 November 2026. Results are attributed to the project’s recorded evaluations; they were not rerun for this portfolio.

NEXT PROJECTProduct Analytics Warehouse

TOOLS BEHIND THE WORK

Explore the stack. Drag the ribbon or select a tool.