🔥 Busiest day: Fri 09 Jan
(forecast ≈ 159, difference vs usual ≈ +49)❄️ Quietest day: Tue 06 Jan
(forecast ≈ 47, difference vs usual ≈ -43)
🟥 4 unusually busy days → plan +1 staff on the busiest flagged days
(Fri 02 Jan, Sun 04 Jan, Fri 09 Jan, Sun 11 Jan)🟦 3 unusually quiet days → reduce prep ~25% on quiet flagged days
(Thu 01 Jan, Sat 03 Jan, Tue 06 Jan)📈 Largest increase vs usual: Fri 09 Jan (≈ +49 demand)
Expected Demand
i.e. transactions, customers, items sold
This heat map shows the model’s expected demand by 2-hour block over the next 3 days.
Each tile gives the expected number of orders in that block. Darker tiles indicate busier periods.
These block-level values are scaled so that, within each day, they sum to the model’s daily expected demand forecast.
Busiest day: Fri 02 Jan (expected daily demand ≈ 143)
Top demand blocks:
- Fri 02 Jan • 12-14 (43)
- Fri 02 Jan • 10-12 (33)
- Sat 03 Jan • 12-14 (33)
Quietest blocks:
- Thu 01 Jan • 16-18 (10)
- Thu 01 Jan • 08-10 (12)
- Suggested actions:
- Add cover or overlap shifts during the busiest blocks
- Increase prep ahead of the busiest morning or midday periods
- Use quieter blocks for breaks, lighter staffing, or prep for next day
Flagged Days (full table)
| Date | Signal | Forecast demand | Likely range (80%) | Usual weekday demand | Difference from usual | Probability busier than usual | Probability quieter than usual | Confidence |
|---|---|---|---|---|---|---|---|---|
| Tue 06 Jan | Unusually quiet | 47 | 37–60 | 90 | -43 | 0% | 98% | Medium |
| Sun 04 Jan | Unusually busy | 99 | 83–121 | 68 | 31 | 92% | 0% | Medium |
| Sun 11 Jan | Unusually busy | 114 | 80–169 | 68 | 46 | 89% | 1% | Low |
| Fri 02 Jan | Unusually busy | 143 | 126–163 | 110 | 33 | 79% | 0% | High |
| Fri 09 Jan | Unusually busy | 159 | 117–219 | 110 | 49 | 79% | 1% | Low |
| Sat 03 Jan | Unusually quiet | 109 | 92–128 | 143 | -34 | 0% | 65% | High |
| Thu 01 Jan | Unusually quiet | 81 | 72–90 | 104 | -23 | 0% | 65% | High |
🌡 Temperature Impact
1°C increase → 6.3% increase in demand
(95% CI: 4.8% – 7.8%)5°C increase → 35.8% increase in demand
(95% CI: 26.6% – 45.5%)
☀ Sunshine Impact
- 10 percentage point increase → 16.8% increase in demand
(95% CI: 15% – 18.6%)
🌧 Rainfall Impact
1 mm increase → 4.5% decrease in demand
(95% CI: -6% – -3%)5 mm increase → 20.5% decrease in demand
(95% CI: -26.5% – -14.1%)
Strong evidence underlying demand ended lower than at the start of the sample
Estimated change from first week to last week: -7.8%
(median: -8%; 95% credible interval: -17.8% to +3.3%)Probability underlying demand is higher than in the first week: 7%
Strong evidence underlying demand was higher than the previous week
Estimated change from previous week to last week: +7.4%
(median: +7.3%; 95% credible interval: -2.6% to +18.3%)Probability underlying demand is higher than last week: 93%
This chart shows underlying demand over time, after adjusting for weekday pattern and weather.
The black line is the estimated demand index, and the shaded band shows the 95% credible interval.
The index is centred at 100, which represents the sample-average underlying demand level. Values above 100 indicate stronger-than-average underlying demand; values below 100 indicate weaker-than-average underlying demand.
At the end of the sample, underlying demand is estimated to be below average.
| Date | Weekday | Signal | Baseline | p50 | p10 | p90 | Lift | P_busy | P_quiet | Confidence | DirConf |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Thu 01 Jan | Thu | Unusually quiet | 104 | 81 | 72 | 90 | -23 | 0.0% | 64.6% | High | Medium |
| Fri 02 Jan | Fri | Unusually busy | 110 | 143 | 126 | 163 | 33 | 79.4% | 0.0% | High | Medium |
| Sat 03 Jan | Sat | Unusually quiet | 143 | 109 | 92 | 128 | -34 | 0.0% | 65.3% | High | Medium |
| Sun 04 Jan | Sun | Unusually busy | 68 | 99 | 83 | 121 | 31 | 91.5% | 0.0% | Medium | High |
| Mon 05 Jan | Mon | Normal | 70 | 71 | 57 | 87 | 0 | 14.1% | 8.4% | Medium | Low |
| Tue 06 Jan | Tue | Unusually quiet | 90 | 47 | 37 | 60 | -43 | 0.0% | 98.4% | Medium | High |
| Wed 07 Jan | Wed | Normal | 90 | 107 | 83 | 141 | 17 | 49.0% | 2.6% | Medium | Low |
| Thu 08 Jan | Thu | Normal | 104 | 94 | 71 | 125 | -11 | 10.0% | 29.5% | Medium | Low |
| Fri 09 Jan | Fri | Unusually busy | 110 | 159 | 117 | 219 | 49 | 79.2% | 1.1% | Low | Medium |
| Sat 10 Jan | Sat | Normal | 143 | 131 | 94 | 188 | -12 | 17.1% | 29.6% | Low | Low |
| Sun 11 Jan | Sun | Unusually busy | 68 | 114 | 80 | 169 | 46 | 88.9% | 0.9% | Low | High |
| Mon 12 Jan | Mon | Normal | 70 | 54 | 37 | 81 | -17 | 8.2% | 55.5% | Low | Low |
| Tue 13 Jan | Tue | Normal | 90 | 93 | 61 | 144 | 3 | 33.3% | 20.8% | Low | Low |
| Wed 14 Jan | Wed | Normal | 90 | 108 | 70 | 173 | 18 | 50.2% | 11.5% | Low | Low |
Signal
Classification for the day:- Unusually busy = demand is likely to be clearly above the usual level
- Normal = no strong signal either way
- Unusually quiet = demand is likely to be clearly below the usual level
Forecast demand
The model’s central forecast for demand on that day.Likely range (80%)
The model’s main forecast range. Demand is expected to fall inside this range about 80% of the time.Usual weekday demand
The typical expected demand for that day of the week, used as the comparison point for the forecast.Difference from usual
The forecast minus usual demand. A positive value means demand is forecast above usual; a negative value means it is forecast below usual. The percentage in the cards shows the size of that difference relative to usual demand.Probability busier than usual
The model’s estimated probability that demand will be unusually high compared with the usual level.Probability quieter than usual
The model’s estimated probability that demand will be unusually low compared with the usual level.Confidence
A simple label based on how wide the forecast range is relative to the central forecast:- High = interval width is less than 35% of the central forecast
- Medium = interval width is less than 60% of the central forecast
- Low = interval width is 60% or more of the central forecast
Category
Menu group being forecast, such as Hot drinks, Pastries or Hot Food.Forecast category demand
The model’s central forecast for demand in that category.Usual category demand
The typical expected demand for that category on that day of the week.Difference from usual
The category forecast minus usual category demand. A positive value means category demand is forecast above usual; a negative value means it is forecast below usual.Probability busier than usual
The model’s estimated probability that demand for that category will be unusually high.Probability quieter than usual
The model’s estimated probability that demand for that category will be unusually low.Confidence
The same confidence measure as above, applied to the category forecast.