Survey CSV cleaning for ratings, comments, and segments
Survey exports mix structured values such as rating and segment with open-text comments. Clean the structured fields without damaging the language respondents entered.
Protect open-text answers
Comments often contain commas, line breaks, quotes, and extra spaces. Those values should remain readable and correctly quoted after cleanup.
- Preview comments after delimiter conversion to confirm columns did not shift.
- Do not strip punctuation from response text unless a downstream system requires it.
- Keep respondent IDs separate from comment text for traceability.
Normalize ratings and segments
Rating fields should be numeric or blank. Segment fields should use one consistent label per group, otherwise charts will split the same audience into multiple categories.
- Turn null-like rating values into blanks before calculating averages.
- Normalize segment capitalization and whitespace.
- Keep a list of invalid ratings for manual review instead of forcing them into a number.
Handle duplicate submissions
Duplicate rows can be caused by exports, retries, or respondents submitting twice. Decide whether the correct key is respondent_id, email, session_id, or a combination of respondent and submitted_at.
- Remove exact duplicate rows when they are clearly export artifacts.
- Review non-exact duplicates because later submissions may be intentional corrections.
- Store a copy of removed duplicates with the cleaned survey file.
Recommended local workflow
- Load the survey sample on the tool page or paste a survey export.
- Keep quote handling, header normalization, null cleanup, and duplicate removal enabled.
- Inspect changed cells in comment, rating, segment, and submitted_at columns.
- Download the cleaned CSV and run a row-count check before analysis.