CITADEL´s Key Performance Indicator Toolbox: Calculating KPIs using Linked Open Data

The main goal of CITADEL is to support Public Administrations in understanding their citizens and in transforming and improving their services by proposing recommendations. CITADEL’s Key Performance Indicator Toolbox is one of the outcomes developed in the context of this project that serves this goal. The toolbox enables Public Administrations (i) to publish data in a reusable format, (ii) to convert data to Linked Data, and (iii) to calculate Key Performance Indicators (KPIs) using the Linked Data.

Introduction

The main goal of CITADEL is to support Public Administrations in understanding their citizens and in transforming and improving their services by proposing recommendations. CITADEL’s Key Performance Indicator Toolbox is one of the outcomes developed in the context of this project that serves this goal. The toolbox enables Public Administrations (i) to publish data in a reusable format, (ii) to convert data to Linked Data, and (iii) to calculate Key Performance Indicators (KPIs) using the Linked Data.CITADEL’s KPI Toolbox affords Public Administrations to find out why citizens stop using public services and helps them to use this information to re-adjust service provision to bring these citizens back in. The KPI Toolbox also helps to identify why citizens are not using a given public service (e.g. because of affordability, low accessibility, lack of knowledge, embarrassment, lack of interest). Where appropriate, Public Administrations can use this information to make public services more attractive and thus heighten the uptake of the services. However, while recommender systems can enhance experiences by providing targeted information, the entry barriers in terms of data acquisition are very high, often limiting recommender solutions to closed systems of user or context models. The CITADEL’s KPI Toolbox developed by imec-IDLab mitigates this by providing an architectural model and a corresponding toolbox that allows harvesting data from various sources. The KPI Toolbox enables Public Administrations to curate datasets that originate from a multitude of formats and to fuse them into semantically enhanced datasets that contain KPIs on e-Government services. Next, the output is processed by an analytics and/or recommender engine - in our case a dashboard - to provide insights on e.g. the usage or demographics of the e-Government services, and to provide suggestions for improvement.

Publishing data on the web

Public Administrations have a lot of useful data and information stored in databases, spreadsheets or other sources. There is a tendency to make some of this data publicly available as Open Data, because Public Administrations have a responsibility to inform their citizens. They might also benefit from applications being developed that use this Open Data. However, it is not easy to publish data from a wide variety of different sources and in different formats in an accessible, uniform way. Tim Berners-Lee, who invented the world wide web (www) 30 years ago, outlined how to tackle this problem on his ‘5 Stars Open Data’ webpage. CITADEL’s KPI Toolbox gives Public Administrations the tools to publish their data as five stars-rated Open Data.

The CITADEL’s KPI Toolbox uses the DataTank to process the data to the three stars-level in an out-of-the-box fashion. The DataTank is a platform to publish data on the web as structured data in an open format (e.g.. json, .csv or .xml). This data can come from different sources such as databases, files, web pages, etc. The DataTank acts a public interface to this data, without the data owner having to export anything or change database systems.

The DataTank can publish data from heterogeneous data sources to structured data on the web

Linked Data

Next, the CITADEL’s KPI Toolbox uses the RML processor to process the three stars-level data to five stars-level data or Linked Data (RDF). This means that everything in the data is assigned a Uniform Resource Identifier (URI) so that it can be referred to worldwide. The RML processor takes data from different sources and converts it to Linked Data using rules that define how to do that. Now the data can easily be linked to other related Linked Data sets, adding context and meaning to the data. The result can then be published again, e.g. by using the DataTank.

Using the RML processor to convert data to Linked Data

Key Performance IndicatorsLinked Open Data allows applications to access and process data in a uniform way, possibly over multiple data sets (this is called federated querying). CITADEL’s KPI Toolbox demonstrates calculating KPIs using Linked Data. It guides the user through a number of steps and, as such, composes the data formula or query dynamically. The KPI Toolbox takes advantage of the fact that Linked Data contains certain semantics and is thus self-explanatory. So there is no need to program or hard-code anything. As a final step the KPI Toolbox plots the results on various charts.

KPIs displayed in the dashboard

To conclude; the KPI Toolbox affords publishing Linked Data that can be used to calculate Key Performance Indicators. In this way, CITADEL’s Key Performance Indicator Toolbox supports Public Administrations in understanding their citizens and in transforming and improving their services.