Barge-port stay predictor
The Port of Rotterdam Authority has developed a dashboard that shows the performance of inland navigation in the past (barge performance monitor). Can that dashboard be extended with a prediction of the port staying time on the basis of publicly available data? The research shows that a rough prediction can be made with public data. To increase accuracy, more data is needed.
On the basis of the available data from the past, it is tested whether the current public data set is sufficient in making an accurate prediction. It turned out that this is not yet the case: more data sets are needed to arrive at more reliable predictions.
Influences on the staying time are: wind speed, day of arrival (which day in the week), but also, for example, the operator concerned and its way of working.
- Prediction of staying times – With public information a rough prediction is possible. But only when aggregated data is added to it, maximum predictability and full insight are possible.
- Extra services – Improved service by The Port of Rotterdam Authority to customers.
- Decision information – More information helps the carriers in making them choose for the most suitable modality.
- Has all possible data that may be helpful been identified? Or have data sets still been overlooked, and what can their possible added value be?
- Are stakeholders prepared to make their data available? (privacy versus transparency)
- How much certainty in the forecast is desirable? How exact does the prediction have to be, to be useful in practice for the various parties involved?
It will be possible to better manage expectations, for example in the choice between transport modalities (inland shipping, road, rail).