Data Bridges: Data Integration for Digital Cities

Melanie Herschel, Ioana Manolescu, Inria Saclay, Université Paris-sud

TL;DR

This paper introduces "DataBridges," an initiative focused on integrating and enriching diverse data sources to create valuable "Open City Data." The goal is to empower citizens and administrations with actionable insights for smarter urban environments, addressing challenges in data acquisition, linking, integration, and visualization.

What Problem Does This Paper Address?

The paper tackles the challenge of fragmented and siloed data within urban environments, often referred to as "Digital Cities." While there's a growing movement towards "Open Data" – making public data freely accessible – simply publishing raw data isn't enough. The problem lies in effectively acquiring, linking, integrating, and making sense of this data to derive meaningful information. This is crucial for developing applications that benefit both citizens (e.g., planning a perfect evening) and city administrations (e.g., resource planning, traffic management). The paper highlights the need for robust methods to bridge these data "islands" and contribute to the vision of Linked Open Data (LOD) in the context of cities.

Methods

The DataBridges activity employs a multi-pronged approach, categorized into three main axes:

  • Producing Open City Data: This involves acquiring data from various heterogeneous formats (relational, text, Excel) and converting it into RDF (Resource Description Framework), a standard for representing linked data. Techniques for data production and best practices are explored.
  • Integrating and Enriching Data: This is the core of DataBridges. It focuses on:
  • Data Linking: Connecting data from structured tables to concepts within ontologies. This allows for deeper analysis by combining data from different sources.
  • RDF Data Integration: Using platforms like RDFGears to define workflows that extend and combine existing RDF datasets, enabling the creation of larger, more comprehensive datasets.
  • Data Quality and Provenance: Developing methods to ensure the quality of the integrated data and track its origin (provenance) to aid in debugging and understanding data transformations.
  • Data Contextualization: Enriching data with context (time, location, topic, agent) to make it more relevant and understandable for specific applications.
  • Exploiting Open Data: This axis focuses on making the integrated data accessible and useful through:
  • RDF Data Management: Developing efficient querying and warehousing solutions for large RDF databases, including cloud-based approaches.
  • User-Centric Applications: Designing intuitive interfaces and applications that leverage visualization techniques and user needs, often incorporating social data analysis. This includes intelligent maps and facetted browsing tools.

Datasets

The paper does not explicitly list specific datasets used for training or evaluation in a formal benchmark setting. However, it mentions the types of data being integrated and exploited, which include:

  • Public administration data (e.g., from Helsinki, French national administration).
  • Social media data (e.g., Twitter datasets).
  • Points of Interest (POI) data.
  • Relational databases.
  • Textual data (e.g., user comments, news articles).
  • Multimedia data (audio, video, text).
  • Geographical data (e.g., from INSEE).
  • Museum and restaurant information.

The paper also refers to the DBPedia ontology as a resource for semantic enrichment.

Key Insights and Ideas

  • Bridging Data Islands is Key: The core idea is that individual Open Data sets are like "islands," and the real value comes from building "bridges" to connect them, fostering a Linked Open Data ecosystem for cities.
  • RDF is the Foundation: The paper consistently emphasizes RDF as the standard format for representing and integrating city data due to its expressive power and flexibility for linked data.
  • Context is Crucial for Relevance: Simply having data isn't enough; enriching it with contextual information (time, location, etc.) is vital for making it truly useful for applications and decision-making.
  • User-Centricity Drives Value: The ultimate goal is to create applications that directly benefit users, whether they are citizens planning their day or administrators making policy decisions. This requires understanding user needs and providing intuitive ways to interact with data.
  • Data Quality and Provenance are Non-Negotiable: For any data-driven application to be trustworthy and effective, ensuring data quality and understanding its origin (provenance) are paramount.

Key Results

The paper focuses on the activity and plans of DataBridges rather than presenting specific, quantifiable results from a single experiment. However, it highlights the following achievements and ongoing work:

  • Development of a multi-partner activity involving universities, research institutes, and companies.
  • Exploration of best practices for producing Open City Data.
  • Development of technologies for RDF data integration (e.g., RDFGears).
  • Work on efficient querying and warehousing of large RDF databases, including cloud-based solutions.
  • Design of user-centric applications such as intelligent maps and facetted browsing tools.
  • Demonstration of an example scenario for planning a "perfect evening" by linking social media and Open Data.

Limitations and Future Work

Limitations:

  • The paper acknowledges that the "exploitation of City Data is still in its infancy" within DataBridges, as the focus has been on building the foundational data infrastructure.
  • The paper does not present a formal benchmark or extensive quantitative results, as it's a position paper describing an ongoing activity.

Future Work:

  • Devising more advanced analytics, applications, and visualizations to extract actual information of interest from the raw data.
  • Investing further efforts in making methods scale to large volumes of data.
  • Extending work to the cloud and defining new algorithms and tools for fully automatic linking, quality assessment, and visualization.
  • Setting up an actual application that combines the technologies developed by activity partners, resembling the introductory example.
  • Further study of C-SPARQL for processing streaming data.
  • Rolling out an experimental semantic data integration platform (ONDA) to integrate legacy relational sources.

Summary updated

Abstract

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et a ̀ la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
BibTeX
@misc{herschel2012data,
  author = {Melanie Herschel and Ioana Manolescu and Inria Saclay and Ioana Manolescu and Inria Saclay and Université Paris-sud},
  title = {{Data Bridges: Data Integration for Digital Cities}},
  year = {2012},
}
Figure