Respostas de inquérito em temas

Question: What one thing would make our meal-kit service better for you?

1. Delivery often arrives after 8pm, I would love a morning slot.
2. Too much plastic packaging in every box.
3. The portions are too small for two hungry adults.
4. Cheaper plans for families please, it adds up fast.
5. More vegetarian recipes, I get bored by week three.
6. The app crashes when I try to skip a week.
7. Please label allergens more clearly on the recipe cards.
8. Deliveries are late about once a month, and nobody tells me.
9. Reduce the packaging, the ice packs and bags are a lot of waste.
10. Prices went up twice this year, it feels expensive now.
11. I wish I could pick my own delivery day and time window.
12. Recipes repeat too often, more variety would be great.
13. Portion sizes vary a lot from week to week.
14. Skipping or pausing a week in the app is confusing.
15. Cheaper per meal if I order more boxes, a discount maybe.
16. Packaging is recyclable which is good, but there is still too much of it.
17. Allergen info is hard to find, I have a nut allergy.
18. More new recipes with fish and less pasta.
19. Everything is fine, I like it as it is.
20. The boxes are too heavy and bulky when they come, the courier struggles, but the food is great.

# Survey analysis: meal-kit feedback

Question: What one thing would make our meal-kit service better for you?
Answers analysed: 20. Tagged answers: 19. Untagged: 1 (#19). Percentages use 20 answers as the base.

## Tagging table
| # | Answer (shortened) | Theme tags |
|---|---|---|
| 1 | Delivery often arrives after 8pm, I would love a morning slot. | Delivery timing |
| 2 | Too much plastic packaging in every box. | Packaging waste |
| 3 | The portions are too small for two hungry adults. | Portion size |
| 4 | Cheaper plans for families please, it adds up fast. | Price |
| 5 | More vegetarian recipes, I get bored by week three. | Recipe variety |
| 6 | The app crashes when I try to skip a week. | App usability |
| 7 | Please label allergens more clearly on the recipe cards. | Allergen labelling |
| 8 | Deliveries are late about once a month, and nobody tells me. | Delivery timing |
| 9 | Reduce the packaging, the ice packs and bags are a lot of waste. | Packaging waste |
| 10 | Prices went up twice this year, it feels expensive now. | Price |
| 11 | I wish I could pick my own delivery day and time window. | Delivery timing |
| 12 | Recipes repeat too often, more variety would be great. | Recipe variety |
| 13 | Portion sizes vary a lot from week to week. | Portion size |
| 14 | Skipping or pausing a week in the app is confusing. | App usability |
| 15 | Cheaper per meal if I order more boxes, a discount maybe. | Price |
| 16 | Packaging is recyclable which is good, but there is still too... | Packaging waste |
| 17 | Allergen info is hard to find, I have a nut allergy. | Allergen labelling |
| 18 | More new recipes with fish and less pasta. | Recipe variety |
| 19 | Everything is fine, I like it as it is. | none |
| 20 | The boxes are too heavy and bulky when they come, the courier... | Packaging waste |

## Themes
| Theme | Mentions | Share | Answers |
|---|---|---|---|
| Packaging waste | 4 | 20% | #2, #9, #16, #20 |
| Price | 3 | 15% | #4, #10, #15 |
| Delivery timing | 3 | 15% | #1, #8, #11 |
| Recipe variety | 3 | 15% | #5, #12, #18 |
| Portion size | 2 | 10% | #3, #13 |
| App usability | 2 | 10% | #6, #14 |
| Allergen labelling | 2 | 10% | #7, #17 |

## Representative quotes
- Packaging waste (4): "Too much plastic packaging in every box." (#2)
- Price (3): "Prices went up twice this year, it feels expensive now." (#10)
- Delivery timing (3): "Deliveries are late about once a month, and nobody tells me." (#8)
- Recipe variety (3): "Recipes repeat too often, more variety would be great." (#12)
- Portion size (2) (fewer than 3 mentions, treat as a signal only): "The portions are too small for two hungry adults." (#3)
- App usability (2) (fewer than 3 mentions, treat as a signal only): "Skipping or pausing a week in the app is confusing." (#14)
- Allergen labelling (2) (fewer than 3 mentions, treat as a signal only): "Allergen info is hard to find, I have a nut allergy." (#17)

## Notes
- Counts come from the tagging table above; an answer may carry more than one tag, so shares can add to more than 100%.
- Themes with fewer than 3 mentions: Portion size, App usability, Allergen labelling.
- Answer #20 mixes packaging with a positive remark about the food; only the packaging complaint was tagged.
- Answer #19 is positive and has no improvement theme.
- 20 answers is a small sample and the survey was not described as random, so do not generalise to all customers.

---
name: survey-analyzer
description: Turns free-text survey answers into themes with counts and representative quotes, using a visible tagging table so every count can be checked. Use when someone has open-ended survey responses and needs to know what people keep saying.
---
# Survey analyzer

Free-text answers in, themes with checkable counts and real quotes out.

## When to use
- The user has open-ended answers (a CSV column, a pasted list, an export from a form tool).
- They want the main themes, how often each came up, and quotes to illustrate them.

## Inputs
- The answers as a file path or pasted text, one answer per line or row. If missing, ask.
- The survey question, so themes are read in context. Ask if it is not given.
- Optional: a theme list the user already has. Otherwise derive themes from the answers.

## Steps
1. Number the answers 1..N in the order given. Keep the original wording; do not edit it.
2. Read all answers once and draft 4 to 10 themes, each a short noun phrase (for example "Packaging waste").
3. Tag every answer with zero, one or more themes. Record the tags in a table: answer number, shortened answer, tags. Answers that fit no theme get "none".
4. Count mentions per theme with a short Python script that reads the tag table, not by eye. Compute share as mentions divided by N.
5. Re-read any "none" answers; add a theme only if two or more of them share a point.
6. Pick one representative quote per theme: a complete answer, copied word for word, with its number.
7. Mark every theme with fewer than 3 mentions as "fewer than 3 mentions, treat as a signal only".
8. Add notes: sample size, answers with mixed content, positive-only answers, and that shares can exceed 100% when answers carry several tags.

## Output
Write `survey-analysis.md` as a new file with: `## Tagging table`, `## Themes` (theme, mentions, share, answer numbers, sorted by count), `## Representative quotes`, `## Notes`. State N and the base used for percentages at the top.

## Rules
- Never invent answers, quotes or counts. Every quote must appear verbatim in the input.
- Write results to a new file, never overwrite the survey data.
- Do not merge or split themes after counting without recounting.
- Do not describe small samples as representative of all customers; say what N is.
- Keep personal details out of quotes, replace names with [name].
- Keep the data local; do not upload it.
- Out of scope: medical, legal, tax or financial conclusions. Flag such answers for a professional.