New paper: Enabling the Use of Sentinel-2 and LiDAR Data for Common Agriculture Policy Funds Assignment

Estrada J, Sánchez H, Hernanz L, Checa MJ, Roman D. Enabling the Use of Sentinel-2 and LiDAR Data for Common Agriculture Policy Funds Assignment. ISPRS International Journal of Geo-Information. 2017; 6(8):255.

  • Abstract: A comprehensive strategy combining remote sensing and field data can be helpful for more effective agriculture management. Satellite data are suitable for monitoring large areas over time, while LiDAR provides specific and accurate data on height and relief. Both types of data can be used for calibration and validation purposes, avoiding field visits and saving useful resources. In this paper, we propose a process for objective and automated identification of agricultural parcel features based on processing and combining Sentinel-2 data (to sense different types of irrigation patterns) and LiDAR data (to detect landscape elements). The proposed process was validated in several use cases in Spain, yielding high accuracy rates in the identification of irrigated areas and landscape elements. An important application example of the work reported in this paper is the European Union (EU) Common Agriculture Policy (CAP) funds assignment service, which would significantly benefit from a more objective and automated process for the identification of irrigated areas and landscape elements, thereby enabling the possibility for the EU to save significant amounts of money yearly.
  • Download paper

New release of DataGraft!

We are delighted to announce the second beta release of the DataGraft platform!

What is DataGraft?

DataGraft serves as the core of the proDataMarket producer portal. DataGraft is an online platform that provides data transformation, publishing and hosting capabilities that aim to simplify the data publishing lifecycle for data workers (i.e., Open Data publishers, Linked Data developers, data scientists).

The DataGraft platform mainly consists of three components – the DataGraft portal, Grafterizer and a cloud-enabled semantic graph database-as-a-service (as shown on the picture below), which is based on a dedicated instance of the Ontotext GraphDB Cloud platform mentioned in a previous blog post.

Main components of DataGraft platform

What’s new?

DataGraft has undergone major changes since the previous version:

  • New asset types for the catalogue and better sharing between users of the platform
    • SPARQL endpoints
    • queries
    • file pages
  • Improved Grafterizer capabilities
    • conditional RDF mappings
    • support various types and formats of tabular inputs
  • Versioning of assets
    • browsing
    • recording of provenance when copying assets
  • Visual browsing of SPARQL endpoints (using RDF Surveyor)
  • New Dashboard
    • more control of user assets
    • instant search and various filters
  • Improved security (authentication) using OAuth2
  • REST API improvements using Swagger
  • Updated version of the semantic graph database, which now supports geospatial queries and serialisation to GeoJSON
  • Various bug fixes and performance improvements
  • Updated user documentation
  • Quota management console allowing users to track their use of resources on the platform

DataGraft beta 2 is available for testing on http://datagraft.io and more details can be found in the platform documentation here. All platform code except the GraphDB Cloud component (used as a service) is open-source and is available on GitHub.

 

New paper: Tabular Data Anomaly Patterns

Sukhobok, N. Nikolov, and D. Roman. Tabular Data Anomaly Patterns. To appear in the proceedings of The 3rd International Conference on Big Data Innovations and Applications (Innovate-Data 2017), 21-23 August 2017, Prague, Czech Republic, IEEE.

  • Abstract: One essential and challenging task in data science is data cleaning — the process of identifying and eliminating data anomalies. Different data types, data domains, data acquisition methods, and final purposes of data cleaning have resulted in different approaches in defining data anomalies in the literature. This paper proposes and describes a set of basic data anomalies in the form of anomaly patterns commonly encountered in tabular data, independently of the data domain, data acquisition technique, or the purpose of data cleaning. This set of anomalies can serve as a valuable basis for developing and enhancing software products that provide general-purpose data cleaning facilities and can provide a basis for comparing different tools aimed to support tabular data cleaning capabilities. Furthermore, this paper introduces a set of corresponding data operations suitable for addressing the identified anomaly patterns and introduces Grafterizer — a software framework that implements those data operations.
  • Download paper

GraphDB Cloud: an on-demand enterprise ready RDF database

We, from Ontotext, are excited to announce GraphDB Cloud – the easy way to introduce you to a semantic database like our signature GraphDB product. The automated tasks in GraphDB Cloud save the organizations the time and effort to install and manage hardware and software as well as the cost to buy it. Compared to a do-it-yourself database, DBaaS offers developers the opportunity to cut down the time it took them to work with their databases and spend their valuable time on creating and innovating instead of administrating.

GraphDB Cloud is one part of the Cognitive Cloud solutions for low-cost and on-demand smart data management.

The users are with the following profile:

  • Small cognitive-technology oriented team in a big organization that needs low upfront and ongoing costs for a database.
  • Start-up companies without a database infrastructure, which requires a reliable technology that scales up along with their business.
  • Corporate solution architects working to solve the challenges their enterprise faces when handling huge amounts of data and information

As a next step, we want to invite you to watch our webinar “GraphDB Cloud – Enterprise Ready RDF Database on Demand” where we introduce you to the DraphDB Cloud console and advise you how you could create custom solutions to address your company’s specific data and information needs.

Understanding territorial distribution of Properties of Managers and Shareholders: a Data-driven Approach

Thanks to the collaboration between Cerved, SINTEF and “Territorio Italia” it was possible to publish a paper which presents a new score developed by Cerved.”Territorio Italia” is an open access peer-reviewed scientific magazine focused on territorial and geographic topics; it is edited by Agenzia dell’Entrate, which is the Italian Revenue authority.

The paper has been announced in the previous blog post. In this post we highlight the main results, the Manager and Shareholders Concentration score and its application to the cities of Turin, Milan and Rome.

Manager and Shareholders Concentration (MSHC) score

The paper introduces the “Manager and Shareholders Concentration (MSHC) score” – an index created with the aim of identifying the wealthiest areas within a certain municipality. This is of
particular interest for the real estate market, especially when there are several wealthy areas within
the same city. The paper thus introduces the index and demonstrate how it can correctly identify
the areas with high real estate values within a city, even when they are located far from the city centre.
The approach proposed in the paper aims to directly observe the distribution of the properties of the wealthiest citizens, who usually choose to move to and live in the most prestigious areas. While this phenomenon can be observed in many cities around the world, in Italy it is particularly evident in the city of Turin: although they are endowed with fascinating city centres, many of the buildings of greatest importance are located on the hills far from the centre. The crucial question becomes to correctly determine which sample of citizens to select and qualify as managers or, more generally, wealthy people. To do this, we used Cerved’s proprietary database – a database containing public data on all Italian companies – to extract information about individuals recognized as shareholders and/or managers. In the context of this work, a shareholder is considered anyone who owns shares above the threshold percentage of 25% of the company’s share capital, while a manager is defined as anyone who holds a key position within a company, accomplishes management duties, and is legally liable for the company’s debts. In calculating the MSHC score, the basic idea is to observe the total number of properties of managers and shareholders per geographic area, comparing this information with the total number of residents in the same geographic area. This approach provides a result that can be immediately visualized graphically using thematic maps; for example, by plotting the score on a map of the city of Turin, it may be noted that the two most relevant areas are, respectively, the centre and the hill on the eastern side of the city.

HEATMAP

The territorial distribution of the MSHC score can be easily observed through a heat map. On the maps, darker colours correspond to high scores, while lighter colours are associated with lower scores. Heat maps also allow the territorial distribution of real estate values to be easily compared, in order to verify whether there is a correlation between prices and scores. For the city of Turin, it was possible to analyse the correlation between the MSHC score and the asking prices for real estate provided by Osservatorio Immobiliare della Città di Torino – OICT (Turin Real Estate Market Observatory), in comparison with their territorial distribution. For the cities of Rome and Milan, the comparison between the MSHC score and real estate values was made using the values published by Osservatorio del Mercato Immobiliare (OMI) of Agenzia dell’Entrate, an important reference for the real estate market on the national level.

TURIN

The score shows high values in the city centre, the hill, and the micro-areas on the western side of the city, while it correctly identifies the south and north areas of the city as less prestigious. This result confirms that the score can also be considered a valuable tool for predicting values on the real estate market.

Figure 1 Territorial distribution of the MSHC score in the city of Turin. The MSHC score is displayed on the map, associating a darker colour with higher scores and brighter colours with lower

ROME

The second city chosen to analyse the MSHC score is Rome, a very complex city due to the vastness of the municipal area that is not comparable to any Italian metropolis, as well as due to the particular shape of some specific areas, namely the proximity to the city-state of the Vatican, the large number of historical and cultural points of interest, and access to the sea.

The size of the Italian capital does not allow the distribution to be observed in detail, but it may be noted that there are more high-value areas, which correspond to actual high-value neighbourhoods and others, which can be defined as emerging neighbourhoods due to the presence of undergrounds and public transit.

Figure 2 Territorial distribution of the MSHC score in the city of Rome. The MSHC score is visualised on the map by associating a darker colour with higher scores, and brighter colours with lower scores

MILANO

The third city used to analyse the MSHC score was Milano – a city that has experienced major changes in recent years. Milan has seen the development of new neighbourhoods and skyscrapers, a universal exposition (EXPO), and a new underground line (with another under development) after years of inactivity. The highest MSHC score is found in the centre of the city, while in the suburbs not many neighbourhoods are identified as particularly wealthy.

Figure 3 Territorial distribution of the MSHC score in the city of Milan. The MSHC score is visualised on the map by associating a darker colour with higher scores, and brighter colours with lower scores

CONCLUSION

The MSHC score illustrated in the paper provides an interesting index that may be used to better comprehend where the richest segments of the population live, and consequently to identify the areas of the city with the highest real estate values. Obviously, although considering this score alone is not enough to support the valuation of real estate property values, together with other indicators under development at Cerved (for real estate valuation) it represents an excellent starting point. For a more in-depth analysis and to observe how much the score is correlated with housing price please have a look at the entire paper and the complete results [1].

References

[1] Stefano Pozzati, Diego Sanvito, Claudio Castelli, Dumitru Roman. Understanding territorial distribution of Properties of Managers and Shareholders: a Data-driven Approach. Territorio Italia 2 (2016), DOI: 10.14609/Ti_2_16_2e

URL to access the article in Italian.

URL to access the article in English.

 

New proDataMarket paper: Combining Sentinel-2 and LiDAR data for objective and automated identification of agricultural parcel features

Combining Sentinel-2 and LiDAR data for objective and automated identification of agricultural parcel features by Jesús Estrada, Héctor Sanchez, Lorena Hernanz, María José Checa and Dumitru Roman

This new proDataMarket paper explains how a comprehensive strategy combining remote sensing and field data can be helpful for more effective agriculture management. Satellite data are suitable for monitoring large areas over time, while LiDAR provides specific and accurate data on height and relief. Both types of data can be used for calibration and validation purposes, avoiding field visits and saving useful resources. In this paper we propose a process for objective and automated identification of agricultural parcel features based on processing and combining Sentinel2 data (to sense different types of irrigation patterns) and LiDAR data (to detect landscape elements). The proposed process was validated in several use cases in Spain, yielding high accuracy rates in the identification of parcel features. An important application example of the work reported in this paper is the European Union (EU) Common Agriculture Policy (CAP) funds assignment service, which would significantly benefit from a more objective and automated process for identification of agricultural parcel features, thereby enabling the possibility for the EU to save significant amounts of money yearly.

Although some issues regarding the generation and improvement of agricultural property datasets were already explained in our previous blog entry (Data workflow in CAPAS), this paper highlights the current results of generation and usage of this new information.

Irrigation patterns map, obtained using Sentinel-2 Process 

The main result of this analysis is how the use of the external, and usually underused, data sources offers a powerful and accurate tool for generating new contrast and validation data for the information used by Spanish CAP Payment Agency, in order to provide a better service to landowners and farmers. As a conclusion, the use of Sentinel-2 series and LiDAR can help to detect areas that are not eligible for grant assignment, support cross-check, and these datasets can be used as a tool for choosing field samples.

The document is available Here.

New Paper: Understanding territorial distribution of Properties of Managers and Shareholders: a Data-driven Approach

Understanding territorial distribution of Properties of Managers and Shareholders: a Data-driven Approach by S. Pozzati, D. Sanvito, C. Castelli, D. Roman, Territorio Italia 2 (2016), DOI: 10.14609/Ti_2_16_2e.

  • Abstract: The analysis and better understanding of the distribution of wealth of individuals in cities can be a precious tool, especially in support of the estimation of real estate values. These analyses can also be used to facilitate decision making in various sectors, such as public administration or the real estate market. In this paper, by making use of publicly available data and of data owned by Cerved, (a credit scoring company in Italy), we can observe the territorial distribution of the properties of managers and shareholders – categories of people usually linked to high economic well-being – and, based on that, we identify the areas of the cities where the value of real estate properties is presumably higher. More specifically, we introduce the Manager and Shareholder Concentration (MSHC) score and validate its accuracy and effectiveness in three Italian cities (Turin, Rome and Milan).
  • Download paper

The proDataMarket Ontology: Enabling Semantic Interoperability of Real Property Data

Real property data (often referred to as real estate, realty, or immovable property data) represent a valuable asset that has the potential to enable innovative services when integrated with related contextual data (e.g., business data). Such services can range from providing evaluation of real estate to reporting on up-to-date information about state-owned properties. Real property data integration is a difficult task primarily due to the heterogeneity and complexity of the real property data, and the lack of generally agreed upon semantic descriptions of the concepts in this domain. The proDataMarket ontology is developed in the project as a key enabler for integration of real property data.

The proDataMarket ontology design and development process followed techniques and design choices supported by existing methodologies, mainly the one proposed by Noy [1]. Requirements are extracted from a set of relevant business cases and competency questions [2] are defined for each business case, so as core concepts and relationships. A conceptual model is then developed based on the requirements mentioned above and international standards including ISO 19152:2012 and European Union’s INSPIRE data specifications. For example, the LADM conceptual model from ISO 19152:2012 is used as reference model to the proDataMarket cadastral domain conceptual model. Afterwards we implemented the conceptual model using RDFS/OWL linked data standard. RDFS is used to model concepts, properties and simple relationships such as rdfs:subClassOf. OWL is built upon RDFS and provides a richer language for web ontology modelling and it is used to model constraints and other advanced relationships, such as the cardinality constraint needed to express the relationship between properties and buildings.

The proDataMarket ontology can be accessed at http://vocabs.datagraft.net/proDataMarket/. The ontology has been divided into several sub-ontologies (see Table below), reflecting the cross-domain nature of the requirements. This modular approach also helped to handle the complexity of the model and made it easier to maintain. In the current version, there are 11 sub-ontologies with 43 native classes and 43 native properties.

Table: Composition of the proDataMarket ontology

Domain/module Namespace prefix URL Classes Properties Business cases
Common prodm-com http://vocabs.datagraft.net/proDataMarket/0.1/Common# 4 4 ALL
Cadaster prodm-cad http://vocabs.datagraft.net/proDataMarket/0.1/Cadastre# 6 16 SoE, RVAS, NNAS, SIM
State of Estate Report prodm-soe http://vocabs.datagraft.net/proDataMarket/0.1/SoE# 4 2 SoE, RVAS
Business Entity Reuse the existing vocabularies, no new classes and properties 0 0 SoE, RVAS
Building Accessibility Reuse the existing vocabularies, no new classes and properties 0 0 SoE
Natural Hazard prodm-nh http://vocabs.datagraft.net/proDataMarket/0.1/NaturalHazard# 1 0 RVAS
Land Parcel Identification System (LPIS) prodm-lpis http://vocabs.datagraft.net/proDataMarket/0.1/LPIS# 1 7 CAPAS
Protected Sites prodm-ps http://vocabs.datagraft.net/proDataMarket/0.1/ProtectedSite# 2 0 CAPAS
Sentinel data prodm-sen http://vocabs.datagraft.net/proDataMarket/0.1/Sentinel# 1 1 CAPAS
Landscape Elements (LiDAR data) prodm-lid http://vocabs.datagraft.net/proDataMarket/0.1/Lidar# 3 0 CAPAS
Assessment prodm-asm http://vocabs.datagraft.net/proDataMarket/0.1/Assessment# 3 3 CAPAS
CensusTract prodm-ct http://vocabs.datagraft.net/proDataMarket/0.1/CensusTract# 1 0 CST,CCRS
Urban Infrastructure prodm-ui http://vocabs.datagraft.net/proDataMarket/0.1/UrbanInfrastructure# 17 10 SIM
Total: 43 43

More than 30 datasets have been published through the DataGraft platform [3] [4] using the proDataMarket ontology as a central reference model. All seven business cases use the proDataMarket ontology in data publishing. More details on the proDataMarket vocabulary can be found in the paper under review: http://www.semantic-web-journal.net/content/prodatamarket-ontology-enabling-semantic-interoperability-real-property-data

References

  • [1] Noy, Natalya F., and Deborah L. McGuinness. “Ontology development 101: A guide to creating your first ontology.” (2001).
  • [2] Grüninger, Michael, and Mark S. Fox. “Methodology for the Design and Evaluation of Ontologies.” (1995).
  • [3] Roman, D., et al. DataGraft: One-Stop-Shop for Open Data Management. 2017. Semantic Web, vol. Preprint, no. Preprint, pp. 1-19, 2017. DOI: 10.3233/SW-170263.
  • [4] Roman, D., et al. DataGraft: Simplifying Open Data Publishing. ESWC (Satellite Events) 2016: 101-106.

Integrating multisectoral datasets: from satellites to real estate scoring model

During a project meeting in Sofia on September 21, 2016, Cerved teamed up with TRAGSA to brainstorm ideas of re-using the TRAGSA methods for processing satellite imagery to analyse green areas in urbanized cities.

Fundamentals of Tragsa Processing

A common feature in Vegetation Spectra is the high contrast observed between the red band and the Near Infrared (NIR) region. The optical instrument carried by Sentinel 2 satellites samples 13 spectral bands, including high resolution bands in the red (bands 4, 5 & 6) as well as bands in the NIR (8 & 8A). Refer to this blog post for more details about processing Sentinel 2 data.

Using the TRAGSA methodology it is possible to isolate and enhance the vegetation, to locate green areas in urban areas. Green areas are important input to the Cerved’s innovative real estate evaluation model (which is being developed within one of the Cerved’s business cases in the project, as introduced in this blog post). Cerved uses open data, to generate indicators of green areas defined for the model: green area coverage and distance to the wood. Operations that Cerved performs to compute these indicators are similar to those that TRAGSA does on satellite data, such as clustering of green areas into big areas and isolating trees and group of trees. This motivated us to experiment with satellite data and TRAGSA’s methodology, to see whether we could potentially use more complete, structured and up-to-date source of green areas information as input to our real estate evaluation model.

Experiment

We identified a highly urbanized Italian city but with particular attention to green areas, which is the city of Turin.

The steps that we followed:

  • extraction of city boundaries of Turin in GeoJSON format by SPAZIODATI
  • selections of good quality imagery for Turin from the Sentinel data repository by TRAGSA
  • processing S2 imagery in order to get a vector layer which indicates the presence or absence of a green area in each pixel (1/0) by TRAGSA
  • display of the green areas of the tiles (see the screenshot below) prototype Amerigo visualisation service, under development by SPAZIODATI
  • data processing and aggregation of the tiles into census cells areas, in order to develop green areas indicators for each census cell, by CERVED
  • integration and testing of the score dedicated to green areas within the business model CCRS (Cerved Cadastral Report Service) by CERVED

image001

The result of this experiment was extremely surprising; the detail and accuracy of this new score in identifying the green areas (not only public green areas) is far greater than accuracy of the other scores, developed on public and open green areas of datasets.

New DataGraft-related papers

DataGraft: One-Stop-Shop for Open Data Management by D. Roman, N. Nikolov, A. Putlier, D. Sukhobok, B. Elvesæter, A. Berre, X. Ye, M. Dimitrov, A. Simov, M. Zarev, R. Moynihan, B. Roberts, I. Berlocher, S. Kim, T. Lee, A. Smith, and T. Heath. Semantic Web journal, 2016.

  • Abstract: This paper introduces DataGraft (https://datagraft.net/) – a cloud-based platform for data transformation and publishing. DataGraft was developed to provide better and easier to use tools for data workers and developers (e.g., open data publishers, linked data developers, data scientists) who consider existing approaches to data transformation, hosting, and access too costly and technically complex. DataGraft offers an integrated, flexible, and reliable cloud-based solution for hosted open data management. Key features include flexible management of data transformations (e.g., interactive creation, execution, sharing, and reuse) and reliable data hosting services. This paper provides an overview of DataGraft focusing on the rationale, key features and components, and evaluation.
  • Download paper

DataGraft: Simplifying Open Data Publishing by D. Roman, M. Dimitrov, N. Nikolov, A. Putlier, D. Sukhobok, B. Elvesæter, A..J. Berre, X. Ye, A. Simov and Y. Petkov. ESWC Demo paper. 2016.

  • Abstract: In this demonstrator we introduce DataGraft – a platform for Open Data management. DataGraft provides data transformation, publishing and hosting capabilities that aim to simplify the data publishing lifecycle for data workers (i.e., Open Data publishers, Linked Data developers, data scientists). This demonstrator highlights the key features of DataGraft by exemplifying a data transformation and publishing use case with property-related data.
  • Download paper

Tabular Data Cleaning and Linked Data Generation with Grafterizer by D. Sukhobok, N. Nikolov, A. Pultier, X. Ye, A..J. Berre, R. Moynihan, B. Roberts, B. Elvesæter, N. Mahasivam and D. Roman. ESWC Demo paper. 2016.

  • Abstract: Over the past several years the amount of published open data has increased significantly. The majority of this is tabular data, that requires powerful and flexible approaches for data cleaning and preparation in order to convert it into Linked Data. This paper introduces Grafterizer – a software framework developed to support data workers and data developers in the process of converting raw tabular data into linked data. Its main components include Grafter, a powerful software library and DSL for data cleaning and RDF-ization, and Grafterizer, a user interface for interactive specification of data transformations along with a back-end for management and execution of data transformations. The proposed demonstration will focus on Grafterizer’s powerful features for data cleaning and RDF-ization in a scenario using data about the risk of failure of transport infrastructure components due to natural hazards.
  • Download paper