Quick Poll / Survey Maker
Create simple polls with multiple question types and visualize results.
关于此工具
Build quick polls and surveys right in your browser. Add multiple questions using single choice, multiple choice, text answer, or 1-5 star rating formats. Switch to preview mode to see the poll exactly as respondents would. Enter responses manually and watch real-time SVG bar charts populate the results view. Export your poll as JSON to save or share it, and import JSON to restore a previous poll. Everything runs client-side — no sign-up required.
使用方法
- 1 Click 'Add Question' and choose a question type.
- 2 Type your question and add answer options (for choice questions).
- 3 Switch to Preview mode to see the respondent view.
- 4 In Simulate mode, enter responses and click 'Record Response'.
- 5 Switch to Results to view bar charts for each question.
- 6 Use Export JSON to save your poll, or Import JSON to restore one.
What this poll maker is built to do
This is a self-contained survey builder that lets you design questions, preview how respondents will see them, simulate answers, and watch results render as live bar charts — all in your browser, with nothing sent to a server. It is best understood as a prototyping and teaching tool: you build the questionnaire, generate sample responses to see how the charts behave, and export the whole thing as a JSON file you can re-import later or hand to a developer. Because there is no backend, it does not collect responses from other people over the internet; instead it gives you a fast, private way to design and test a poll before you commit it to a hosted platform.
The four question types
| Type | Respondent sees | Best for |
|---|---|---|
| Single Choice | Radio buttons — pick exactly one | Mutually exclusive options (yes/no, one favorite) |
| Multiple Choice | Checkboxes — pick any number | "Select all that apply" questions |
| Text Answer | A free-text box | Open feedback and short comments |
| Star Rating | A 1–5 star scale | Satisfaction and quality scores |
Choice questions start with two options and let you add up to six, and you can never delete below two — a sensible floor, since a choice with one option isn't a choice. Each option's label is editable inline.
The four modes, in order
- Edit — add, reorder, retype, and configure questions.
- Preview — see the form exactly as a respondent would, with real radios, checkboxes, text areas, and interactive stars (this view is for looks; it doesn't record anything).
- Simulate — actually fill in the form and click Record Response to store a submission. Repeat to build up a sample dataset.
- Results — aggregates every recorded response into charts.
How results are calculated
The results view tallies each question according to its type, which is worth understanding so you read the charts correctly:
- Single and multiple choice are shown as horizontal bar charts with each option's count and its percentage of total respondents. For multiple-choice, note that percentages can sum to more than 100% — that is expected, because one person can tick several boxes.
- Star rating shows the distribution across 1–5 stars plus an average. The average is the standard weighted mean: multiply each star value by how many people chose it, sum those, and divide by the number of ratings. Three people giving 5, 4, and 3 stars yields
(5+4+3)/3 = 4.00. - Text answers are listed verbatim as a bullet list, since free text can't be averaged.
The bars are drawn as inline SVG and scaled so the most-chosen option fills the track, which keeps small differences visible even when the absolute counts are low.
Saving and sharing your design
Export JSON downloads a file containing both your questions and any recorded responses. Import JSON reads that file back in, restoring the full state. This is the tool's portability mechanism: keep a master copy of your survey, version it, or pass the structure to whoever will deploy the live version. Since everything is JSON, it is also easy to inspect or edit by hand.
Practical tips for writing a good poll
- Use single choice unless multiple genuinely applies. Checkboxes invite people to select everything, which dilutes your signal; reserve them for true "all that apply" cases.
- Keep option lists short. The six-option cap is also good survey design — long lists cause fatigue and random clicking.
- Pair a rating with a text box. A star score tells you how people feel; a follow-up text question tells you why.
- Test edge cases in Simulate. Record a few responses where someone skips a question or picks nothing, then check Results to confirm the charts handle it the way you expect before going live.
- Phrase questions neutrally. Leading wording ("How much did you love our service?") biases answers; the tool can't fix wording, only you can.
What it is not
Because there is no shared backend, this tool will not collect anonymous responses from a public link, enforce one-vote-per-person, or store results between visits beyond the JSON you export. It is a design, demonstration, and analysis sandbox — fast, private, and free — not a hosted voting service. For everything up to the point of public distribution, that is exactly what you want.