Log-Based Sheet Format

Harmonize works best when your sheet is structured as a log: one header row and rows of data underneath. We use the headers to map your messages to the right columns.

Basic structure

  • Row 1 = headers. Each column has a clear name (e.g. Date, Crop, Amount, Block, Notes).
  • Row 2 onward = data. Each new message we log becomes one new row. We don’t overwrite existing rows.

We read your sheet’s structure (and sometimes sample data) to understand what kind of data you’re logging. That helps us parse messages like "Harvested 50 lbs tomatoes from Block A" into the right columns.

Example: Harvest log

| Date       | Crop     | Amount | Block  | Notes   |
|------------|----------|--------|--------|---------|
| 2025-02-15 | Tomatoes | 50 lbs | Block A|         |
| 2025-02-16 | Eggs     | 2 doz  |        | Sold $8 |

With headers like this, a message such as "Harvested 30 lbs squash from Block B" can be parsed into Date (today), Crop, Amount, Block, and we append a row.

Multiple tabs

You can connect one tab per "dataset" in the dashboard. Each tab can have its own set of headers and its own log. When you connect a sheet, you choose which tab to use for that connection. We only write to the tab you selected.

Best practices

  • Keep headers clear and consistent. Names like "Date", "Crop", "Amount", "Block" work well. Avoid merging cells in the header row.
  • Don’t put extra text or instructions in the first row. Row 1 should be only column names.
  • Use one theme per tab. A "Harvest" tab with Date, Crop, Amount; a "Sales" tab with Date, Item, Quantity, Revenue—each tab is one log type.

When you connect a sheet, we may analyze the headers (and optionally sample rows) to build a small "data model" so the parser and query system know what fields exist and how to interpret your messages.

Queries and your structure

Questions you send (e.g. "How many harvests this week?", "Total sales yesterday?") are answered using the data in your connected sheet(s). The query system uses the same header and data structure to aggregate and return results. So a consistent, log-based layout improves both logging and query answers.

Related