Data Interlinking and Validation

SPARQL-Integrate

SPARQL-Integrate is a swiss army knife for realizing small to medium sized data integration workflows as a mere sequence of SPARQL queries stored in a .sparql fileIntegrate heterogeneous data with standard SPARQL syntax plus function extensions.

The tool uses the plugin system of Apache Jena’s SPARQL engine (ARQ) for adding the functionality to access local and remote data and to process JSON, CSV and XML formats. Introduces SPARQL functions which compute a single RDF term from its arguments, e.g. json json:parse(string) and SPARQL property functions (can be seen as magic RDF properties) which transform literals into multiple SPARQL result set rows, e.g. { ?jsonLiteral json:unnest ?itemVar }

Trip Update Interface

Trips are the fundamental entity for analyzing mobility patterns in a Smart City. How many trips are done by car every week? How many in bike? Can we encourage the usage of greener transportation modes?

The trip Update Interface allows citizens to contribute with trips they have made, and to verify/amend trips detected automatically, for example with the transport mode detector component

Data Analytics

Transport Mode Detector (TMD)

The Transport Mode Detector is a component that takes a timestamped GPS and accelerometer trace of a citizen and returns (i) the trips made by the citizen, and (ii) a classification of the transport modes of each leg of the trip. The ML techniques were used in combination with Data Mining strategies to pre-process the streaming data and evaluate the models to provide confident labels for specific trips like bike, bus, car, train and walk. The result is a model able to predict new data with high accuracy and high confidence level.

TomTom Parking Probabilities

This analytics module developed by TomTom takes data on parking spots (on-street and off-street, for bikes, motorbikes and cars) and historical trip data for these type of vehicles, and produces a map of parking probabilities

Data Storage and Fusion

QROWD-DB

QROWD-DB is the main storage component facilitates the discovery and exploitation of data in specific processes and services.
It is comprised by three main elements:
1) The Linked Data Storage: its main goal is to store and manage linked data, entities;
2) The Big Data Storage: dedicated to the storage of the personal big data generated by mobile phone of citizens;
3) The Business Logic: a set of software components previously developed by the University of Trento that allows the partners to access the underlying two
systems and databases

Link-Discovery and Data-Fusion (LDDF) utils

QROWD includes a set of utils for discovering links and fusion data in RDF. The utils take as input several data sources on the same entities (e.g., location of mobility infrastructure accroding to different sources, including crowdsourcing) and allow the creation of clusters of sameAs entities, that can then be fusioned according to trust scores on the the original data sources.