The BuiltHub Platform

The dashboards are conceived as a user-friendly central platform area where end users can download and analyse data.

In further detail, taking the Emissions dashboard as an example, predefined categories allow easy data filtering by country, group of countries such as EU27, year, indicator (such as CO2 emissions), and sector. Multiple choices are possible, choosing for example several years or countries. A table and charts show the respective data. A geographic map with the selected countries and the data source, the Emissions Database for Global Atmospheric Research (EDGAR), are shown as well. The tabular data can be downloaded as a CSV (Comma-Separated Values) file. For the overall visual appearance of the Emissions dashboard, see Figure 4. The other dashboards have a similar look-and-feel.

The Energy dashboard, derived from H2020 HotMaps data, allows filtering by end use (space heating, space cooling, water heating, lighting and electrical appliances, etc.), sector (residential or services), building type depending on sector (e.g., single-family, apartment blocks, offices), and by country and year.

The Census dashboard offers information about the number of buildings and dwellings, or non-residential uses taken from the EU Census Hub. It allows filtering by construction period and demonstrates filtering by NUTS (Nomenclature of Units for Territorial Statistics) level, i.e., geographical unit, from national (NUTS 0) to small-region (NUTS 3) level.

Several other dashboards showcase more advanced services for testing and evaluation by pioneer users of the platform. For example, the Comparison dashboard showcases the comparison of two different datasets (IEE ENTRANZE and H2020 HotMaps) for the same indicator. This allows end users to evaluate the adequacy of a dataset and to choose the most appropriate one for their purposes. The Geo Information dashboard demonstrates the possibility to visualise layered data on a geographical map. Not only single values for a NUTS territorial unit can be shown but also a continuous distribution of values such as, solar irradiance.

To further simplify the data analysis and knowledge transfer for an end user, storylines have been created combining different datasets that inform about a topic through predefined graphs and explanatory text. The graphs are interactive, by hovering with the mouse over parts of the graph, the respective information is shown.

The first storyline regards and delves into the energy consumption of the EU countries in the buildings sector. It shows various interactive charts like the one in Figure 5.

The second storyline showcases the transformation into knowledge at district level of a regional dataset, for example for the Flemish province of West Flanders, courtesy of the BE REEL! Project. The dataset reports the number of households at municipal level by building type (terraced, semi-detached, detached, apartments, other) and construction period. It further indicates the energy consumption reduction potential offered by energy retrofitting.

The third storyline shows energy consumption over the years by country and end use, and at different NUTS levels. This is a powerful feature, as it allows monitoring and evaluating the energy efficiency improvements (and therefore renovation efforts) over time down to NUTS3 regions. By selecting multiple regions, they can be directly compared and represented in different formats (table, map, various types of charts) readily available for download.

The fourth storyline presents a machine learning (ML) analysis carried out to support the Swedish long-term renovation strategy. It can be used to gain main insights about the methods used and results obtained. While this analysis is too complex to be performed on-the-fly, it shows the possibility offered by an expert service that can be activated for interested data providers such as governments from the national to the local level.

After the topics and the storylines, a third section in the dashboard area refers to the available datasets and choosing appropriate ML methods for analysing the datasets. In this section, the BuiltHub Data Catalogue reports all datasets included in the database and gives a description of them. The last item in this section is a ML method mapper, which shows the appropriate ML methods for a specific type of dataset based on scientific literature, this is shown in Figure 6.

While the dashboards are considered to be the main arena of the platform for most users, the data library is a part of the platform dedicated to users who need more detail and customisation. The data library offers datasets exploration, targeted querying, and upload. It is subdivided into the following sections:

• Ontology Viewer
• Data Catalogue
• SPARQL editor
• Upload dataset

Ontology viewer

The platform uses the WebVOWL viewer, which is a web-based tool for interactively visualising the BuiltHub datasets as objects with their properties and vocabulary. The user can consult both technical data (creators, date, etc.) and statistical information related to the BuiltHub ontology, querying specific information about the different datasets and their interrelations. The viewer also provides the possibility of exporting the ontology in different formats.

Data catalogue

This section provides a list of the datasets available on the platform and a description of each dataset. By selecting a dataset, the user receives detailed information on structure, metadata, and predicates of the dataset. For example, metadata is offered on geographical coverage NUTS levels, indicators, units, time range, periodicity, and SIEC (Standard international energy product classification). Predicates allow for machine-machine interaction.

SPARQL editor

This editor allows users to write dedicated queries in the SPARQL query language to select and combine data from all BuiltHub datasets. A series of premade queries and the Data Catalogue support the user in combining, comparing, and analysing the data.

Upload dataset

This section consists of a simple form allowing the user to select any of the available datasets to expand them by uploading files in the suitable structure and format from their own device to the BuiltHub platform database.

In addition to the manual data upload feature, SPARQL query facility, and data/charts download options on the dashboards, the BuiltHub platform offers data import and export integration services allowing users to establish connections with external BI (Business Intelligence) and IDE (Integrated Development Environment) software.

The first integration service is the OData feed. This open data protocol works with most modern BI tools, such as MS Excel and Power BI. In this way, users can establish a connection between the BuiltHub platform and their BI software to load the desired datasets.

The second integration service is the REST API integration. Data can simply be pulled by the respective GET request specifying the URL of the dataset.

Finally, a query created and saved in the SPARQL editor can be called from any BI software to load the data corresponding to the query in the BI software. As for the second integration service, to do this, it is sufficient to specify the URL corresponding to the query.

The Help section provides several written guides as well as video tutorials on how to use the different sections on the platform.