Aline Senart’s research while affiliated with SAP Research and other places

Ensuring data quality in Linked Open Data is a complex process as it consists of structured information supported by models, ontologies and vocabularies and contains queryable endpoints and links. In this paper, the authors first propose an objective assessment framework for Linked Data quality. The authors build upon previous efforts that have identified potential quality issues but focus only on objective quality indicators that can measured regardless on the underlying use case. Secondly, the authors present an extensible quality measurement tool that helps on one hand data owners to rate the quality of their datasets, and on the other hand data consumers to choose their data sources from a ranked set. The authors evaluate this tool by measuring the quality of the LOD cloud. The results demonstrate that the general state of the datasets needs attention as they mostly have low completeness, provenance, licensing and comprehensibility quality scores.

Ensuring data quality in Linked Open Data is a complex process as it consists of structured information supported by models, ontologies and vocabularies and contains queryable endpoints and links. In this paper, the authors first propose an objective assessment framework for Linked Data quality. The authors build upon previous efforts that have identified potential quality issues but focus only on objective quality indicators that can measured regardless on the underlying use case. Secondly, the authors present an extensible quality measurement tool that helps on one hand data owners to rate the quality of their datasets, and on the other hand data consumers to choose their data sources from a ranked set. The authors evaluate this tool by measuring the quality of the LOD cloud. The results demonstrate that the general state of the datasets needs attention as they mostly have low completeness, provenance, licensing and comprehensibility quality scores.

The Open Data movement triggered an unprecedented amount of data published in a wide range of domains. Governments and corporations around the world are encouraged to publish, share, use and integrate Open Data. There are many areas where one can see the added value of Open Data, from transparency and self-empowerment to improving efficiency, effectiveness and decision making. This growing amount of data requires rich metadata in order to reach its full potential. This meta-data enables dataset discovery, understanding, integration and maintenance. Data portals, which are considered to be datasets' access points, offer metadata represented in different and heterogenous models. In this paper, we first conduct a unique and comprehensive survey of seven meta-data models: CKAN, DKAN, Public Open Data, Socrata, VoID, DCAT and Schema.org. Next, we propose a Harmonized Dataset modeL (HDL) based on this survey. We describe use cases that show the benefits of providing rich metadata to enable dataset discovery, search and spam detection.

Linked Open Data (LOD) has emerged as one of the largest collections of interlinked datasets on the web. In order to benefit from this mine of data, one needs to access to descriptive information about each dataset (or metadata). This information can be used to delay data en-tropy, enhance dataset discovery, exploration and reuse as well as helping data portal administrators in detecting and eliminating spam. However, such metadata information is currently very limited to a few data portals where they are usually provided manually, thus being often incomplete and inconsistent in terms of quality. To address these issues, we propose a scalable automatic approach for extracting, validating, correcting and generating descriptive linked dataset profiles. This approach applies several techniques in order to check the validity of the metadata provided and to generate descriptive and statistical information for a particular dataset or for an entire data portal.

Linked Open Data (LOD) has emerged as one of the largest collections of interlinked datasets on the web. In order to benefit from this mine of data, one needs to access descriptive information about each dataset (or metadata). However, the heterogeneous nature of data sources reflects directly on the data quality as these sources often contain inconsistent as well as misinterpreted and incomplete metadata information. Considering the significant variation in size, the languages used and the freshness of the data, one realizes that finding useful datasets without prior knowledge is increasingly complicated. We have developed Roomba, a tool that enables to validate, correct and generate dataset metadata. In this paper, we present the results of running this tool on parts of the LOD cloud accessible via the datahub.io API. The results demonstrate that the general state of the datasets needs more attention as most of them suffers from bad quality metadata and lacking some informative metrics that are needed to facilitate dataset search. We also show that the automatic corrections done by Roomba increase the overall quality of the datasets metadata and we highlight the need for manual efforts to correct some important missing information.

Linked Open Data (LOD) has emerged as one of the largest collections of interlinked datasets on the web. In order to benefit from this mine of data, one needs to access to descriptive information about each dataset (or metadata). This information can be used to delay data entropy, enhance dataset discovery, exploration and reuse as well as helping data portal administrators in detecting and eliminating spam. However, such metadata information is currently very limited to a few data portals where they are usually provided manually, thus being often incomplete and inconsistent in terms of quality. To address these issues, we propose a scalable automatic approach for extracting, validating, correcting and generating descriptive linked dataset profiles. This approach applies several techniques in order to check the validity of the metadata provided and to generate descriptive and statistical information for a particular dataset or for an entire data portal.

The Internet has created a paradigm shift in how we consume and disseminate information. Data nowadays is spread over heterogeneous silos of archived and live data. People willingly share data on social media by posting news, views, presentations, pictures and videos. SNARC is a service that uses semantic web technology and combines services available on the web to aggregate social news. SNARC brings live and archived information to the user that is directly related to his active page. The key advantage is an instantaneous access to complementary information without the need to dig for it. Information appears when it is relevant enabling the user to focus on what is really important.

The increasing size and availability of web data make data quality a core challenge in many applications. Principles of data quality are recognized as essential to ensure that data fit for their intended use in operations, decision-making, and planning. However, with the rise of the Semantic Web, new data quality issues appear and require deeper consideration. In this paper, we propose to extend the data quality principles to the context of Semantic Web. Based on our extensive industrial experience in data integration, we identify five main classes suited for data quality in Semantic Web. For each class, we list the principles that are involved at all stages of the data management process. Following these principles will provide a sound basis for better decision-making within organizations and will maximize long-term data integration and interoperability.

With today’s public data sets containing billions of data items, more and more companies are looking to integrate external data with their traditional enterprise data to improve business intelligence analysis. These distributed data sources however exhibit heterogeneous data formats and terminologies and may require helping the user merging data coming from heterogeneous sources. remix is a Business Intelligence (BI) solution that offers business users a productive environment to easily create highly formatted reports they would ultimately like to see. Via rich visual context-aware interactions, users can quickly combine shared data sets and reuse report parts. This enhanced collaboration combined with the provision of a multi-source semantic layer give users the power to make more effective and informed decisions on virtually any relevant data source or BI resource wherever they are.

With today's public data sets containing billions of data items, more and more companies are looking to integrate external data with their traditional enterprise data to improve business intelligence analysis. These distributed data sources however exhibit heterogeneous data formats and terminologies and may contain noisy data. In this paper, we present RUBIX, a novel framework that enables business users to semi-automatically perform data integration on potentially noisy tabular data. This framework offers an extension to Google Refine with novel schema matching algorithms leveraging Freebase rich types. First experiments show that using Linked Data to map cell values with instances and column headers with types improves significantly the quality of the matching results and therefore should lead to more informed decisions.