Breakout Group on Policy Databases



Hello all!
I am new to openMod (silently following the mailing list for a while now) and will attend the Frankfurt workshop, and would like to introduce myself a little!
My name is Martin, I work for DLR on agent-based modelling of electricity markets; currently I investigate the uptake of PV battery systems.
During my work I realized that policy instruments should be more readily available for modelling exercises, so I proposed a breakout group on energy policy open data, hope that kind of initiative is okay!?

Looking forward to April


Dear Martin,

your kind of initiative is warmly welcome!
Since we are working on a collaborative database for open energy data we are happy to get ideas about policy instruments. The oedb is a PostgreSQL database with a web interface.


Interesting Subtopics:

  • Data Sources (Primary & Secundary sources)

  • Abstraction & Characterization

  • Is a (relational) database the right structure


CO2 Reduction Targets:
Structured data on INDCs:

RES Targets and Expansion:

Policy databases:

TODO: Find out Licenses and Terms of Use
Does a oedb mirror make sense?


Outcome of our discussion in the data abstraction sub-group:

Policy is particularly challenging to store in a highly structured database. Policy instruments evolve both in details and structure very often, which would put a strain on the schema update requirements of a highly structured schema.

Another point that was identified during the discussion is the relative higher priority of national-level policy, relative to say household-level policy (e.g. appliances rating etc.), which led us to recommend to focus first on national level data, then build out from there.

Our recommendation regarding structure is a hybrid nature: Entries should contain some structural information, as well as free text information. A first draft schema proposal has been formulated:

  • Geographical scope: One or several (e.g. Europe, Germany, Berlin)
  • Technology scope: One or several (e.g. Wind, RES)
  • Date enacted: When the law/regularion/target/etc. was published
  • Date valid: When does it come into effect?
  • Date of expiry: When will it no longer be in effect?
  • Characterization: What type of policy instrument is this? E.g. Subsidy, tax, renewables target, court ruling, etc.
  • Policy text: Free text containing the text of the law/regulation/court ruling/etc.
  • Binding or non-binding: Whether or not this policy instrument is binding or not (since some policy texts describe in tent but not a binding decision)
  • Source: Traceable source of this entry

This would allow policy entries to be searchable and listed in say reverse chronological order, but still require some human interpretation before being compiled and input into a model. Given the high heterogeneity of models and modelling questions, we see this as an appropriate first solution.


After a close look, it seems IEA has a very useful “policies and measures” database (linked above in Martin’s post.
The database operates as an interactive map. After clicking on a country, clicking on “All Policy Information” will take you to a list of policies and measures placed on a timeline, with contextual data in the background as a graph. The list can be filtered by keyword and date range. Each list entry contains a relatively rich amount of structured information as well as the text of the policy or the measure.

The drawback of the IEA database is that the data is not downloadable or exportable.


Here’s some documentation from the discussion in the database/tech-sub-group:

Generally the group agreed that energy modeller’s needs for concrete policy-information (RE-targets, emission targets, support schemes, auction volumes etc.) has to be determined in the first step. There should be a google-doc to collect these modelling needs.

The group discussed that there are certain pieces of policy-information that can be quantified and usually stored in relational database but policy processes in general don’t function in strict relations that can be represented in a relational database. Therefore any policy-database for energy-modellers could be in a two-data-bases-manner: One where quantifiable data is put (relational db) and one where texts, laws, regulations, court-decisions, comments and other documents can be stored and “losely associated” with tags/keywords and/or each other. A good example of this is the way the Open Energy Data Base (OEDB) and the Open Energy Platform (OEP) work. Maybe the non-relational part could also be displayed in the form of a wiki. In the OEDB a schema called “policy” should be added.

General remark:
A challenge for the process of setting up a database for policy-information will be that the structure of laws change while the setting-up-process takes place thus necessitating a dynamic and flexible structure (and also design-process).


  • There is the recommendation to license databases with the Open Data Base License (ODBL).
  • How are laws and regulatory publications and respective information within those licensed? Is it legal to autmatically read them out?

Next Steps:
(1) Put up a google-docs file for modeller’s expectations towards the policy-database with a deadline to fill it out (which deadline? who sets it up where?)
(2) Structure the needs and expectations and hand them to Ludvig who will pass them to “database-specialists” who can give a recommendation on how (if) these expectations could be represented in a relational database
(3) Set up a beta-version of the policy-database and receive feedback from working modellers

Future Outlook:

  • How can quantifiable information (structured information) automatically be extracted from law texts and regulations?


Just for notice, I updated the openmod wiki with the links from Martin’s post so that visitors who come across this site, don’t need to search the forum for policy data.