Paste or upload a CSV of your network - demand points with weights - and watch a live twin run it: demand pulses tick by tick, a hub serves it, and service and utilisation move in real time. The full-featured twin lives at controltower2.html; this one runs on whatever dataset you bring.
| Point | Lat | Lng | Demand |
|---|
A digital twin is a live model fed by real data - not a static picture. Here the data is the twin: every row of your CSV becomes a breathing demand point, and the model only exists because your dataset does.
Your CSV becomes demand points on the map. The twin places a hub at the demand-weighted centroid, then runs 100 stochastic ticks. Each tick every point’s demand fluctuates around its base and the hub has only capacity units to give away - so the allocation policy decides who actually gets them. Colours are per point: green served, amber partial, red starved.
The twin starts you at 100% of mean demand - and that already fails, because half of all ticks land above the mean. Size capacity to the demand tail, not the mean. But every unused unit is money parked, so nudge capacity down until utilisation earns its keep. Then switch the policy: the fill rate never moves (you can only ship what you have) but who suffers changes completely - that is a management decision, not a maths one.
