Cold Chain Overview
Live conditions and active hazards, read straight from the Open-Meteo and GDACS APIs. For shipment scoring, go to Shipment Prediction.
Current conditions
Loading from Open-Meteo…
Next 48 hours
Temperature and relative humidity at the selected location. Two panels on a shared time axis — the two measures do not share a scale.
Temperature °C
Relative humidity %
Source: Open-Meteo Forecast API
Seven-day outlook
Daily highs and lows, rainfall and peak wind
| Day | Conditions | High / low | Rain | Wind |
|---|
Source: Open-Meteo Forecast API
Ambient exposure, next 7 days
Hours the outside air sits in each band. Useful for judging how hard a reefer or a passive box has to work on the ground.
Derived from Open-Meteo hourly temperature and humidity
Active hazards here
GDACS events affecting the selected country
Source: GDACS event list API
Global disaster alerts
Current GDACS events worldwide. Orange and red alerts are the ones that close roads, ports and airports.
| Alert | Type | Event | Country | Severity | Started |
|---|
Source: GDACS (Global Disaster Alert and Coordination System), gdacs.org
Shipment Prediction
Describe the shipment, then press Predict. The model runs in your browser; nothing is sent anywhere.
Information
How to read the tool and what the risk bands mean. Column-by-column definitions live on Model & Dataset.
How to use it
Four steps
- Open Shipment Prediction and pick Country From, Country To and Food Type. Start typing and choose from the list — typing
indoffers India and Indonesia. - Set transit time, number of legs, package type and fill ratio.
- Press Predict. Nothing is scored until you do, so you can change your mind freely while filling the form.
- Read the risk band, the shelf-life check for your food type, then What if to see how a shorter trip, fewer hand-offs or other packaging changes it.
The risk model only uses four inputs known before shipping. Sensor readings such as temperature and humidity are only known after the trip, so they are not model inputs — for live weather and hazards, use the Overview page, which reads Open-Meteo and GDACS directly.
Risk bands
Model & Dataset
Dataset
shipment-sensor-dataset.csv
Model
XGBoost, gradient-boosted trees
Test-set metrics
Variables in the dataset
Descriptions marked inferred come from the column name. Please confirm them against your data dictionary.
| Variable | Role | Meaning | Mean | Range |
|---|
ROC curve
Precision-recall curve
Calibration
Predicted risk vs actual failure rate, 10 bins. Points on the dashed line mean the percentages can be trusted.
Confusion matrix
Threshold trade-off
Lower thresholds catch more failures but flag more shipments
| Threshold | Flagged | Precision | Recall |
|---|
What the model relies on
Share of the model's total gain
Door-opens models
Used for the handling forecast and what-if
About Us
The team behind ColdChain Monitor.