Skip to main content
Talk to us

Dashboards · Key idea

Ask your records a question with a query

Write a question in a few words, such as kilos delivered by buyer this quarter, and see the answer as a chart and a table.

Every report in Analytics → Reports is a query. To find, save, share or add reports to a dashboard, see Reports library; this article explains the query itself.

In Reports you ask your own records a question in a short, plain language. A query names where the data comes from, what to show and how to group it:

FROM deliveries
SHOW sum(quantity) IN kg AS delivered
BY recipient
DURING this_quarter
COMPARE TO previous_period

That reads: from deliveries, show the kilos delivered by each buyer this quarter, compared with the quarter before. Bags and tonnes are converted to kilograms for you.

How a query is built

  • FROM names a dataset: lots, deliveries, certificates, sites, field work and more.
  • SHOW says what to show: totals such as count(), sum(quantity) or avg(moisture), or plain fields to get a list of records.
  • BY groups the totals: by a field (a buyer, a product) or by day, week, month, quarter or year.
  • WHERE keeps only some records: status = "active", product IN ("Café arábica"), NOT HAS certificate.
  • SINCE, DURING and COMPARE TO set the period and what it is compared with.
  • VISUALIZE picks the card: bar, line, donut, histogram, table and others. On a report page, see Choose how it is drawn.

Press Ctrl+Space for suggestions and Ctrl+Enter to run. Names you type, such as a buyer, are matched to your directory. If something is not right, the query says what and how to fix it, in words.

Build it, save it, put it on a dashboard

The Builder tab writes the same query with menus: what to show, how to group, which records and the period. Each change appears in the query text, and each change to the text appears in the builder. Save keeps the query for you, or for the whole company if you manage dashboards; saving it again keeps the earlier versions. Add to dashboard puts it on one of your dashboards as a card that follows the saved query and takes the dashboard's period when the query has none. Everyone who opens the card sees what their own role allows.

Forest results of your plots

The sites dataset also has each plot's forest result, the same one the Forest tile shows on a lot's Origin tab. Sites that are not plots of land, such as warehouses, have none.

  • forest_result: no_loss (no forest loss detected after 31 Dec 2020), review (forest loss detected: review) or not_enough_data.
  • loss_in_forest_ha: the hectares of tree cover loss after 2020 on land that was forest in 2020.
  • forest_computed_at: when the result was computed.
  • reviewed: whether someone recorded a review of the result.
  • review_current: whether the latest review was made on the figures the plot has now.

For example, to count your plots by result:

FROM sites
SHOW count()
BY forest_result

The result describes the data. The due-diligence conclusion is the company’s.

Where the numbers come from

Every number comes from your company's own records, read with your role's permissions. Rows that rest on the ledger say so; directory and review records say "Platform records". A query only reads: it never changes a record. People whose role does not show personal data cannot filter or group by it.