Chrono Hub

The Chrono Hub represents an open web platform that aims to support 6ST applications (WSB, Campsite, Oxfam) that require capabilities to predict the future state of transportation 'entities and artifacts' (e.g. from Oxfam collection vehicles to sensor-rich road-side gullys) based on historic and real-time information (e.g. GPS traces, gully silt levels).

Chrono Hub


We developed the Chrono Hub, as a cloud-based service to support the development of predictive applications by providing a platform for sharing and execution of prediction algorithms. Mobile applications can utilise the hub by registering real-world objects and accessing predictions of the future state of these objects, whilst algorithm developers can benefit from access to a larger dataset for refining their predictions.
By exploring the design of an open platform that allows developers to store, access and process historic, real-time and prediction information the Chrono Hub can offer several key benefits:

  • The ability to store and manage future data.
  • Access to shared data from other applications.
  • Minimise development time for new prediction applications.
  • Give prediction model developers rich datasets to use.
  • Allow access to the latest prediction models.
  • Strengthen predictions with external datasets.
  • Use cloud technology to store large datasets and process them quickly.


phone-exploded.pngThe Chrono Hub was developed alongside many of the Sixth Sense Transport applications. As we developed the applications we learnt more about their requirements for predictive data.
The majority of applications in this area need to be able to predict future events and states of entities. The hub allows application developers to be able to upload data representing a particular entity (sensor, mobile application, user) and attach associated meta-data. Changes in the state of the object can be attached to an entity and these events determine the state of the object at a certain point in time.

A user can upload custom prediction algorithms and attach them to entities. The prediction models use the meta-data and events attached to the entity to predict the future state of that entity. Once this information has been predicted the hub stores the future data so that it can be accessed by the user's applications as well as used for more predictions in the future.
A comprehensive set of RESTful application programming interfaces (REST APIs) provides access to all entities and events, and allows users to upload and manage prediction algorithms as well as linking these to entities. A Publish-Subscribe Service allows third-party applications to subscribe to predictions: new events generated as a result of a prediction are pushed out to all subscribers automatically.



To gain early insights into the feasibility of cloud support for predictive mobile applications we used the Chrono Hub to support several applications:

In the goWSB application parents use the predicted movement of the Walking School Bus to help coordinate their morning activities. The goWSB application pushes the current position of the bus to the hub, a script then runs to generate predictions of the arrival time for each stop based on historical data from the days before. These predictions are then disseminated to all of the parents' applications through subscriptions to the hub.

The original design of the hub has been deployed in the Sustainable Transport in the Wild project that has enabled extensive trials "in the wild" to support highways maintenance applications.

Key Findings

Our experiences of using the Chrono Hub to support a number of mobile applications have given us new insights into how predictive cloud services should be designed and the implications for the developers of future mobile applications.

Unexpected Usage

Predictive applications and prediction models are much more interested in the state of the world rather than a single entity. This requires some mechanism for grouping entities together and using the combined data from these feeds to predict the future state of the world rather than a single entity. Currently, users of our hub have to download the prediction feeds for multiple entities and then combine them themselves.

Supporting An Open Prediction API

We predict that developers will require scope for parameterisation of prediction models, allowing a model to be tailored to better suit a particular application need. For example, support may be provided to request that an algorithm takes an optimistic or pessimistic view of the world, or to include or exclude a particular data source.

Open Debugging Tools

Prediction Model developers need an environment in the hub where they can test their Prediction Model and view the warning or error logs if it is not working correctly. Problems have occurred as prediction models are developed locally where there are discrepancies between the hub and the local environment.

Prediction Uncertainty

Screen-Shot-2014-11-29-at-15.55.48.jpgPredictions of the future states of entities naturally have a degree of uncertainty. It is important that, developers are able to assess the degree of confidence of any given prediction and have the ability to relay this information to their users. This mockup shows how predictions might be represented to users in a way they can understand, this can help users make better decisions from the information they receive in applications.

It is often difficult to represent the change in the state of multiple assets effectively and highlight future opportunities to a user. Providing appropriate options to allow users to customise their experience and only be notified about the assets they care about is important.



Design of the Chrono Hub


We understand the extent to which behavioural change in transport habits and practices can be facilitated through the creation of a new form of ‘transport network’, based on extending social tablet.pngnetworking principles to transport users.

Read More

Mobile Apps

The project has developed a suite of mobile phone apps for each of the corresponding research contexts. Watch videos and read details of the projects aims, phone-exploded.pngkey findings and outputs.

Read More


The 6ST team comprised researchers from the universities of Southampton, Edinburgh, Salford, people.pngBournemouth and Lancaster.

Read More