Contamination Alerts and the Weekly Digest

Sporeline writes two kinds of AI insight for your farm on a schedule. Once a day it checks your contamination rate and writes a Contamination Risk alert when something needs attention, and every Monday it writes a Weekly Digest of how the week went. You’ll find both under Analytics > AI Insights, and the digest is also emailed to the farm owner. There’s nothing to set up.

Both use the same numbers as the contamination figure on your dashboard, so the alert, the digest and the dashboard agree with each other.

How the Contamination Rate Is Worked Out

Contamination rate = contaminations recorded in the period ÷ units started in the period, on the farm’s calendar. Units are blocks, or lot units on a lot-tracked farm.

A few things follow from that:

• A contamination counts on the day you record it, whether your farm discards contaminated blocks or keeps them in the Contaminated stage. Under the default setting a flagged block is discarded, and it still counts.
• Days are your farm’s days, in the time zone set for your farm, so a block recorded late in the evening counts on that day and not the next.
• If your farm tracks lots, the units your lots started with and the units you threw out as contaminated are added in, and the insight says “units” instead of “blocks”.
• If nothing was started in a period, there’s no rate for it. Sporeline doesn’t call that 0%.
• A contamination can come from a block started before the period, so a short period can hold more contaminations than starts. A rate over 100% wouldn’t mean anything, so there’s no rate for that period either, just as on your dashboard.

The Contamination Risk Alert

Each day Sporeline works out the rate over the same days as your dashboard’s contamination figure, today and the 30 days before it, and over the 31 days before that. It writes an alert only when:

• the rate for the recent period is above 5%, or
• the rate has gone up by more than 3 percentage points since the period before, when that period has a rate to compare with.

When the recent period has no rate (nothing started, or more contaminations than starts), no alert is written.

The alert lists the contaminations by type and by recipe, so you can see where they’re coming from. When the rate is within the normal range, no alert is written.

The Weekly Digest

The digest covers the 7 full days before the day it’s written, so Monday’s digest covers Monday through Sunday of the week before, and compares them with the 7 days before that. The day it’s written isn’t included, since it has barely started. For each of the two weeks it gives:

• blocks (or units) started
• contaminations recorded and the contamination rate, or a note that the rate isn’t meaningful when more contaminations were recorded than blocks started
• harvests and the yield in pounds

For this week it also gives how many blocks finished, completed or discarded.

The digest also shows two current totals, which are not this week’s activity: your active blocks right now, and your blocks by stage right now. These count every block on the farm today, however long ago it was started, and the digest labels them that way.

The numbers behind each digest are saved with it, so the week’s figures and the current totals stay as they were on the day it was written.

When You Won’t Get an Insight

• A farm with no block history yet gets no digest and no alert.
• A farm whose daily counts haven’t been built yet gets no digest and no alert until they are. Sporeline records this for us to fix, so it never shows you a 0% rate it didn’t really work out.
• If the AI service is down when the job runs, that day’s alert or that week’s digest is skipped and the next run carries on as normal.

If Analytics Won’t Load

Analytics loads your farm’s data into the browser in a few steps, and if one of them fails the page tells you which one: asking for your data, preparing it, downloading one of its files, or starting the analytics engine. Press Retry to load it again, and if the engine is the step that failed, another browser may work. Each failure is also recorded for us, so if you send feedback from that page we can see what went wrong.

Related Pages

• AI Insights: where to find and dismiss your insights
• KPIs: the numbers behind your operation
• Data Health: find and fill gaps in your records