With the tremendous amount of data needed to understand urban freight, the effort to collect them is equally as heavy. Wouldn’t it be great if the bulk of the data to be collected could be automated? It would save a lot of time, increase reliability (i.e. reduce human error in collection), increase the intensity of data collected, and potentially reduce the cost!

As mentioned in the previous lesson, some data collection activities are already being automated, and many more are being converted. And although some of the surveys mentioned could already be automated, the level of implementation globally still remains low.

Below are several remarks about the automating several of the data types mentioned above.

Direct measurements and observations at establishments

By monitoring the establishments, we might be able to collect information on the number of freight trips the establishment generates. Imagine a loading area with restricted access, where only designated trucks or drivers are able to access with an electronic key. The actual number of times the specific driver visits can be directly counted because of the electronic records. If this is further linked to the type of goods or retail outlet, we get very detailed information regarding the duration and schedule of each visit. At the operational level, improvements to internally regulate delivery time-windows can be done. At the traffic level, we know how much freight traffic is generated at the particular commercial area or center. The MobileDock System introduced in Week 2 can also provide this information through the loading bay booking system.

However, monitoring goods flows, in terms of weight or volume, may be difficult. If the orders are made electronically and the shipper is willing to open up this source of data, this data could be collected. In practice, this is difficult and unlikely. The same is true for knowing “levels of inventory” at an establishment.

Direct measurements and observations on road

New technology for automatic weighing of vehicles using weigh-in-motion systems have been developed, but its use is not yet widespread. Such technology are developed to help with enforcement of vehicle weight limits but have the potential to help collect data on the weight of goods carried on vehicles. This data might be coupled with other vehicle ownership databases to guess the types of goods carried.

But, even without this new technology, understanding traffic caused by freight vehicles may also be sufficient. Traffic sensors built into the road can detect types of vehicles and automatic road pricing gantries can collect more detailed information at the boundaries of the city. These technology are already quite mature and are used for monitoring freight and informing transport planning.

Records of routing using telematics

Many fleet management systems record GPS locations and other on-board indicators (such as fuel efficiency), which may be shared with researchers. Apps that provide services to vehicle drivers might also request in return the location data. Such data can be rich! The dwelling time at a particular company can be measured, which gives information about the loading and unloading time for the transport operation.

However, fleet companies and drivers themselves may not want to share this data with any third-party, even if they were researchers. Some drivers use the vehicles also for private matters, which therefore becomes a privacy issue. Another question is whether the data transmitted will be “matched” to trip purpose, which also exposes the customers.

Big data

In line with the digitalization of cities, as well as the promotion of open data, many “other data” streams have already been automated. Its function, however, still depends on the development, testing and validation of data analytic techniques, which are not yet widely exploited in urban freight research. Big data is still a very large concept, which houses different types of data, objectives and approaches. What we are mostly thinking about are big data analytics, which is the way to derive insights or forecasts based on analyzing large, diverse and fast data streams. In general, these are NOT theory-driven analyses, which is both its weakness and strength. So, while there are some use cases of the big data techniques, the applications are still not yet significant, in terms of informing decisions.