Most of this site describes what ExcelInsight does. This page shows it. Below is the actual output of the analyser over a real 156-row spreadsheet — the column types it inferred, the statistics it computed, the gaps it found and the charts it proposed. These numbers are generated when the site is built, not written by hand, so they are exactly what you would get by loading the same file yourself.
sample-sales-data.csv — a product order export with 156 rows and 9 columns, covering January to September 2026. It is synthetic, but shaped like a real export: regions and sales channels are unevenly distributed, prices vary within each product's band, and 5 orders have no sales representative recorded.
No column mapping, no configuration. Types are detected from the values themselves — and the statistics below are computed for every numeric column automatically.
| Column | Detected type | Distinct | Missing | Computed statistics |
|---|---|---|---|---|
order_date | Date | 119 | — | — |
region | Categorical | 4 | — | top value “North America” (59 rows) |
product | Categorical | 8 | — | top value “Headset” (28 rows) |
category | Categorical | 3 | — | top value “Accessories” (75 rows) |
channel | Categorical | 3 | — | top value “Online” (66 rows) |
sales_rep | Categorical | 6 | 5 | top value “A. Okafor” (36 rows) |
units | Numeric | 36 | — | min 1.00 · max 40.00 · mean 12.65 · median 8.00 |
unit_price | Numeric | 156 | — | min 29.02 · max 1,871.70 · mean 222.20 · median 115.38 |
revenue | Numeric | 156 | — | min 93.13 · max 20,195.04 · mean 1,954.51 · median 1,146.98 |
Every figure here is a KPI the tool derives on its own — sum, count, mean — and each one recalculates when you apply a filter in the live dashboard.
Because revenue was detected as summable and region and category as categorical, these groupings are proposed without being asked for.
order_date was recognised as a date column, so the rows are bucketed by month automatically — total units by month.
Gaps are reported rather than quietly filled in:
sales_rep — 5 of 156 rows empty (96.8% complete)From 9 columns the analyser derived 30 candidate charts. You start from these rather than from an empty canvas, and keep the ones that are useful. A sample:
Everything above is static. Open the same dataset in the live tool to drag the tiles around, filter it, add your own charts and export a PDF.