public-comment-summarizer

Public Comment Summarizer

Purpose: Convert raw comments into an honest, usable summary for decision makers, while preserving the official record. A summary never replaces the underlying comments, which must stay in the record.

Required inputs

• All comment sources: oral testimony (recordings/minutes), emails, web forms, letters, petitions, survey responses, social comments if included
• The question the body is deciding and the adopted criteria (if quasi-judicial)
• Comment period dates; identifying fields (address, ward, role) if collected

Workflow

1. Inventory and de-duplicate. Count unique commenters, form letters, and petitions separately. Record the timing of each comment against the deadline.
2. Code each comment by position (support / oppose / mixed / question) and by theme (traffic, cost, safety, equity, environment, process, property values, design, and others emerging). Allow multi-coding.
3. Report counts with caution. Public comment is self-selected and not a poll. State this, avoid percentages implying representativeness, and show who is and is not heard (renters, non-English speakers, young people); compare commenter geography to affected area where possible.
4. Extract substance: specific facts, data, alternatives, and conditions proposed. For quasi-judicial matters, separate comments relevant to the adopted criteria from comments about general preference; tell the decision maker which are record evidence.
5. Flag misconceptions with the correct, sourced fact and a suggested neutral response (feeds faq-builder).
6. List unanswered questions and owner/answer-by date.
7. Present representative quotes (short, accurate, attributed per policy; avoid exposing private details).
8. Offer response options: modify the proposal, add a condition, extend the comment period, hold a second session, publish a response memo, or no change with reasons.
9. Document how comments were considered; keep the raw set and the summary together under records retention.

Output structure

• Method and counts, with caveats
• Top themes ranked by frequency and by consequence
• Position map by area/group
• Misconceptions and answers
• Unanswered questions
• Recommended responses

Example

Theme: "Traffic safety on Elm" (41 of 128 comments; 9 offered specific data). Misconception: "The plan removes all parking" - the plan removes 12 of 140 spaces (staff report p.6).

Red flags / escalate to the city attorney

• Comments containing threats, personal data, or claims against the city
• Late comments in a closed-record quasi-judicial hearing
• Potential evidence of bias or ex parte communication by a decision maker

Common mistakes

• Treating volume as consensus
• Omitting minority views
• Dropping comments that criticize staff or officials (viewpoint neutrality)

Not legal advice.