Turn Raw Social Comments Into a Sentiment Report

Data Analysis Claude intermediate

Clean, classify and theme a scrape of social comments into an honest sentiment report with reply recommendations.

When to use it: Use after a campaign or a busy stretch when you have a pile of social comments and want a defensible read, not a vibe.
You are a social-listening analyst for an Australian small business. Turn the raw comments below into a sentiment report the owner can trust — cleaned first, counted honestly, quoted exactly, caveated properly.

<context>
COMMENTS: [PASTE THEM — one per line, with platform and the post/context each came from; usernames optional]
PERIOD / CAMPAIGN: [WHAT THIS COVERS — e.g. launch week for the new range]
BRAND TERMS: [YOUR NAMES/HANDLES/PRODUCT NAMES, SO MENTIONS ARE READ RIGHT]
WHAT PROMPTED THIS: [THE WORRY OR HOPE — e.g. did the price change land badly?]
</context>

Before analysing, clean the batch: exclude spam, bot-patterned, and off-topic comments — report how many were excluded and why, so the denominator is honest.

<task>
1. Classify each retained comment: sentiment (positive / negative / mixed / neutral) — sarcasm and ambiguity flagged 'uncertain' rather than guessed — plus theme tags built FROM the data, not imposed.
2. Stats framed honestly: 'of the N comments analysed', never 'customers think'. Include the uncertain count.
3. Per-platform split if multiple platforms are present — tone differs by platform and mixing them blurs the read.
4. Theme table: theme | count | sentiment lean | 1-2 exact verbatims (quoted, unedited) | bearing on the stated worry/hope.
5. Influence note: if any commenter appears repeatedly or visibly carries reach, name the effect on the read — 10 comments from one person is one opinion with stamina.
6. Reply recommendations: which 3-5 comments most deserve a response (questions, fixable complaints, glowing posts worth amplifying) with a drafted reply each in the business's voice — thanks specific, criticism acknowledged without arguing, problem-solving taken to DMs.
7. Caveats, stated plainly: commenters are not a sample of customers; silence is data you don't have; a report on N comments guides tone and follow-ups, not strategy.
</task>

<output_format>
Method note (excluded count) → stats → per-platform split → theme table → influence note → reply drafts → caveats.
</output_format>

Rules: no invented or edited quotes; no percentages without their base shown; if fewer than ~15 usable comments remain, say the batch supports replies but not conclusions. En-AU spelling.

Copy the block above straight into Claude — anything in [BRACKETS] is yours to fill in.

Want it tuned to your business? Bring it to the free weekly call and we'll adapt it live.

Join the free call

More data analysis prompts

Decide Which Customer Questions Your Data Can Actually Answer

Triage your customer-behaviour questions against the data you really hold, and get spreadsheet-level recipes for the answerable ones.

Match the Right Simple Analysis to Your Business Data

Pick analysis techniques that fit your question, data shape and sample size — with spreadsheet steps and misread warnings.

Use Sales and Feedback Data to Back Your Best Offerings

Combine sales, margin and feedback evidence into a keep/grow/fix/kill call on each product or service.