Urban Water Demand Forecast

Water · production

Production models
3
Top accuracy
96.4%
Runs / day
1,284
All models
Accuracy
94.8%
Total runs
1,108
Latency p95
147 ms
Drift score
0.42
About

Hourly DMA-level demand prediction blending weather, tourism and Ramadan calendars for storage planning.

Probabilistic
Edge-deployable
Bilingual
Explainable
OT-isolated
Inputs
Tune parameters before the next inference run.
Input parameters
24h

How far ahead the model projects load.

90%

Width of the predictive band.

15min

How often inference re-runs.

Geographic scope for inference.

Feature toggles
Inference summary
Horizon24 h
Confidence90%
Refresh15 min
Regiondhofar
Resolution15min
Active features3 / 5

Changes apply to the next scheduled inference window.

Results
Charts update live as you tweak the inputs above.
Accuracy trend
Drift gauge
0.42
within bounds
Daily throughput (k rows)
Run outcome mix
success 11warning 3failed 2
Runs
Recent inference jobs and their outcomes.
Run IDStartedDurationRowsAccuracyStatus
water-demand-forecast-r10002030-06-20 03:148m 10s124,00094.9%
failed
water-demand-forecast-r10012030-06-19 03:149m 11s125,32095.43%
success
water-demand-forecast-r10022030-06-18 03:1410m 12s126,64095.15%
success
water-demand-forecast-r10032030-06-17 03:1411m 13s127,96090.56%
success
water-demand-forecast-r10042030-06-16 03:1412m 14s129,28090.67%
warning
water-demand-forecast-r10052030-06-15 03:148m 15s130,60093.46%
success
water-demand-forecast-r10062030-06-14 03:149m 16s131,92092.94%
success
water-demand-forecast-r10072030-06-13 03:1410m 17s133,24093.46%
failed
water-demand-forecast-r10082030-06-12 03:1411m 18s134,56093.44%
warning
water-demand-forecast-r10092030-06-11 03:1412m 19s135,88090.39%
success
water-demand-forecast-r10102030-06-10 03:148m 20s137,20090.68%
success
water-demand-forecast-r10112030-06-09 03:149m 21s138,52093.95%
success
water-demand-forecast-r10122030-06-08 03:1410m 22s139,84092.95%
warning
water-demand-forecast-r10132030-06-07 03:1411m 23s141,16090.53%
success
water-demand-forecast-r10142030-06-06 03:1412m 24s142,48093.06%
failed
water-demand-forecast-r10152030-06-05 03:148m 25s143,80095.42%
success
Lineage
Sources → Features → Model → Consumers.
SCADA telemetry
Weather feed
AMI readings
CRM tickets
Feature store · v2.4
Training pipeline · airflow:dag-411
Eval harness · v1.8
Control-room dashboard
Outage triage agent
Billing reconciler