Comparing current metrics against past baselines reveals seasonal regressions—like a 40% checkout drop compared to the same hour last Tuesday.
SignalFlow provides .timeshift(offset) to shift any stream forward in time so past behavior aligns directly with the current timestamp. Common offsets include '1h' (hour-over-hour), '1d' (day-over-day), and '7d' or '1w' (week-over-week).
Once aligned, you can subtract the shifted baseline from the current stream or divide to compute a percentage change: ((current - last_week) / last_week) * 100. Plotting both the current series and the percentage delta makes seasonal regressions immediately obvious.
now = data('orders.count', rollup='sum').sum()
last_week = now.timeshift('7d')
pct_change = ((now - last_week) / last_week) * 100
pct_change.publish(label='WoW Orders % Delta')