Performance
In-browser performance benchmark
How results are checked
Named statistical review pending. Selected automated reference cases are published on the validation page.Choose 10,000 to 250,000 synthetic respondents and run the benchmark. Your browser generates the data, parses it as CSV and computes crosstabs with chi-square, then shows how long each step took on this device.
Run the benchmark on this device
Synthetic data is generated in this tab. Nothing is uploaded and nothing is sent anywhere.
Benchmark
What is measured
- Synthetic survey rows are generated in this tab with a fixed seed: demographics, a brand grid, NPS, satisfaction drivers, just-about-right scales and flavor ratings. No real respondents are involved.
- The rows are written to CSV text and then parsed with the same CSV parser the workspace uses, including number detection and import-limit checks.
- Three crosstabs are computed with the same engine the workspace uses, each with Pearson chi-square, association measures and column-proportions tests.
- Each step is timed with performance.now(). Pauses between chunks, which keep the page responsive, are not counted.
How to read the results
Timings depend on the device, browser, other open tabs, power settings and whether the browser has warmed up its JavaScript engine; running the same size twice often gives different numbers. Compare runs on the same device rather than across devices.
This benchmark runs on the page itself so the timings are easy to observe. In the workspace, imports and tables run in background workers where the browser supports them, and real files differ from synthetic data in width, text length and category counts. Workload limits beyond this benchmark are documented on the capacity page.
Next steps