ePolicy was an FP7 project that focused on regional planning and promoted the assessment of economic, social and environmental impacts during the policy-making process, at both the global and individual levels. For the latter, ePolicy aimed to calculate social impacts through opinion mining of e-participation data extracted from the web. To aid policy makers, citizens and stakeholders, ePolicy relied heavily on visualisation tools providing easy access to data, impacts and political choices.

The ePolicy project aimed to integrate two tightly connected modelling viewpoints: the global (i.e. regional) and the individual perspectives. The first considers global objectives, financial aspects, constraints, and large-scale impact assessment on the economy, society and the environment. The second identifies individual needs, merges attitudes and reactions to specific political decisions, and defines the best policy implementation strategy.

All the technical work packages were converged into a flexible open-source architecture.

The techniques used were general and could be applied at different political levels: regional, provincial, urban planning, and both public and private projects.

Policy Context

ePolicy was an FP7 STREP project partially funded by the European Commission under the ICT theme, Objective 5.6 ICT Solutions for Governance and Policy Modelling. It aimed to support policy makers in their decision-making processes through a multi-disciplinary effort to engineer the policy-making life cycle.

The ePolicy project ran from October 2011 to September 2014. The consortium consisted of research institutes from Italy, Ireland, Germany, the UK and Portugal. Three implementation partners from Italy were responsible for disseminating the research outcomes. They validated the system using the Regional Energy Plan (REP) of the Italian Region of Emilia-Romagna. as a test project. 2.6 million Euro of the total budget of 3.3 million was funded by the European Commission.

Aims

The ePolicy project aimed to:

• support policy makers in their decision-making process through a multi-disciplinary effort to engineer the policy-making life cycle, integrating global (i.e. regional) and individual perspectives on the decision-making process;
• evaluate the economic, social and environmental impacts of policy making, at both global and individual levels;
• derive social impacts through opinion mining of e-participation data;
• aid policy makers, citizens and stakeholders using visualisation tools.

The expected direct outcomes of the project were:

• a flexible tool for optimisation and decision support for policy making at a global level, taking into account objectives, constraints, financial issues, and impacts on the environment, the economy and society;
• at an individual level, an agent-based simulation approach to identify the best policy implementation strategies;
• a game-theoretical approach to model the interaction between the global and the individual levels;
• opinion mining techniques to reveal social impacts derived from e-participation data;
• a novel application of visual analytics techniques to support policy makers in the decision-making process and help citizens and stakeholders by providing a more informed evaluation;
• to integrate the components described above, an open source tool that is accessible and re-usable in other policy contexts;
• extensive activities to disseminate the project's results and prepare for the exploitation of the over-all solution as well as its individual components.

The expected social and economical benefits of the project were:

• improved prediction of policy impacts leading to more efficient implementation of regional policies and better identification of the benefits and consequences for citizens and businesses;
• increased engagement of citizens and wider use of ICT tools, resulting in innovative interactions between citizens and government;
• improved transparency on the impact of economic decisions on society;
• improved capacity to react to societal challenges, and increased trust in government by stakeholders and the public at large.

Description of target users and groups

The ePolicy project focused on regional planning. It promoted the assessment of economic, social and environmental impacts during the policy-making process, at both the global and individual levels. For the individual aspects, ePolicy aimed to derive social impacts through opinion mining of e-participation data extracted from the web. To aid policy makers, citizens and stakeholders, ePolicy relied heavily on visualisation tools providing easy access to data, impacts and political choices.

Description of the way to implement the initiative

Model

The ePolicy project aimed to integrate two tightly connected modelling viewpoints: the global (i.e. regional) and the individual perspectives. The first considers global objectives, financial aspects, constraints, and large-scale impact assessment on the economy, society and the environment. The second identifies individual needs, merges attitudes and reactions to specific political decisions, and defines the best policy implementation strategy.

The enabling technologies were:

• optimisation and decision support for the global planning viewpoint,
• agent-based simulation for the individual perspective, and
• game theory to regulate the interaction between these two levels.

A number of underlying technologies were required to create an open, transparent and accountable tool:

• E-participation tools and opinion mining technologies were used to extract social attitudes.
• Opinion mining identified social impacts that should be considered at both global and individual levels. At the global level, opinion mining produces ex ante (i.e. expected) policy evaluation: citizens' opinions collected from selected websites indicate the general sentiment on specific policy fields.
• After the global optimisation component has produced a set of alternative (feasible and optimal) scenarios, citizens and stakeholders can express their ex post (i.e. evaluated) opinions on the policy options available, producing expectations for each alternative.
• Advanced visualisation techniques help policy makers — who are generally not IT experts — in the decision-making process. They also assist citizens and stakeholders in providing informed opinions on the policy options.

All the technical work packages were merged into a flexible open-source architecture.

Technology solution

The policy model in the ePolicy system is a multi-criteria combinatorial optimisation problem, says Michela Milano, Professor at the University of Bologna. For instance, in the field of energy politics it might be used to merge environmental impacts and constraints, territory-based constraints, existing energy plants and generated power, and the budget of the public agency. The model produces a set of alternative scenarios, based around shares of different energy sources and energy efficiency measures in a future enrgy mix, together with the Strategic Environmental Assessment for each of these choices.

The agent-based simulation is a model that captures the behaviour of classes of citizens (clustered according to age, education, gender, number of family members, income and environmental sensitivity). Putting together agents and making them interact, one can observe an emerging social behaviour that can be interpreted as the social reaction to specific policy instruments.

For example, if you provide incentives for the production of energy from renewable sources, you can simulate how many people will actually adopt this technology, and that way assess the impact on the energy market.

Finally, the somewhat conflicting interests of the region on the one hand and the individual on the other are handled using a game-theoretical approach. So the two components representing these interests, namely the simulator and the optimisation component, have to reach a compromise — an equilibrium in which both goals are satisfied as well as possible. This is accomplished through a series of interactions between the two components.

Opinion mining

Opinion mining is one of the components of the ePolicy Decision Support System (DSS), says Luís Torgo, Associate Professor at the Department of Computer Science of the University of Porto, and lead developer of the opinion mining component. The main goal of this component is to provide policy makers with an overview of how citizens feel about certain topics related to the decision domain.

While the outcome of this module was initially planned to be used as input to the optimisation modules, the consortium has changed its role as a result of the research carried out during the project. In effect, it was decided that the outcome of this module was more useful as a 'side source' of information that could be directly consulted by policy makers. It allows them to adjust their decisions based on various aspects of the alternative policies that were fed into the other components of the ePolicy system.

In the specific context of this project, the main goal of the opinion mining module was to provide information on trends in public sentiment regarding various aspects related to energy policies. This information is automatically inferred by the module from the interaction of citizens in a predefined list of e-participation forums.

The opinion mining module is an integral part of the ePolicy system, but it can also be used as a stand-alone solution in other application domains.

The interactions between the main system and the opinion mining component are twofold, Milano adds. First, the component interacts with the visualisation tools as opinions are visualised and aggregated at various granularities. Second, the components' aggregated output is used in the decision support system as a representation of the public sentiment toward a particular policy-making choice.

Technology choice: Mainly (or only) open standards, Open source software

Main results, benefits and impacts

The proposed techniques are general and can be applied at different political levels, such as regional, provincial, urban planning, and both public and private projects. In the first instance, however, the ePolicy project focused on regional planning in the Emilia-Romagna Region of Italy, whose Regional Energy Plan (REP) served as the case study.

The case study aimed to provide a tool to support regional planners in creating an energy plan in line with strategic EU and national objectives, and consistent with financial and territorial constraints. The plan was to include the results of opinion mining. It was to be well assessed from an environmental perspective, and optimal with respect to one or more metrics.

In addition to the regional plan, ePolicy aimed to provide a portfolio of implementation instruments (fiscal incentives, tax exemptions and investment grants) to push the energy market, and society in general, in the direction envisaged by the plan.

Visualisation

The system helps policy makers by suggesting alternative policy plans, assessing their environmental sustainability, and proposing corresponding implementation instruments, says Milano. In addition, the plans are visualised together with the sentiments learned from recent discussions on energy-related topics. This visualisation helps policy makers compare plans and get in touch with public opinion.

Return on investment

The ePolicy project has yielded a complete policy-making platform tailored to aiding policy makers in the field of renewable and sustainable energy, Milano continues. The platform has been used by the Emilia-Romagna Region during the project, by the Piemonte Region afterwards, and has been presented to a number of other government bodies.

Individual components have also been deployed. The module to optimise plans has been used by Sis-Ter, a consultancy company in Imola, to work on municipality plans. It has also been used by the Eulabor Institute, a non-profit in water sustainability development, to assess maritime policies and their environmental impact.

Benefits

According to Milano, the ePolicy platform has — in general — improved the work of policy makers in three ways:

• they can now measure the impact of policies with indicators that can be shared with other actors and stakeholders;
• the ePolicy platform presents alternative scenarios, whose economic and environmental footprints can be compared;
• through opinion mining, policy makers can get a feeling for the impact of their choices on society at large.

Track record of sharing

We have developed a methodology for exporting the results from ePolicy to other policy domains, says Milano. For example, we have applied the ePolicy system to the Regional Operational Program that allocates regional budgets in Italy. We are currently working on an application at municipal level, and in other areas such as maritime policies.

Lessons learnt

One important lesson learnt is the advantage of close integration between simulation and decision support, Milano says. That allows for simulated outcomes that are based on social behaviour and trends, as processed by the optimisation components. We recently published the paper 'Empirical decision model learning' on this topic in the journal 'Artificial Intelligence'.

Scope: National, European, Regional (sub-national)