The Open Contracting team met in Berlin at the Open Knowledge Festival on July 15-17, to discuss the Open Contracting project and the data standard. We presented the Open Contracting: Disclosing Data, Engaging for Results session on Wednesday afternoon. Four breakout groups formed within the last 30 minutes of the Open Contracting session to discuss the following pertinent topics: Governance, Standard, Demand and Tools.
Governance (Lead: Jose Alonso)
Participants: Andrew Lamb (Appropedia Foundation), Maggie Murphy (Transparency International)
The governing principles are at the core of the standard. Leading examples and the levels of engagement are key. The standard needs to remain independent and have multistakeholders facilitate the discourse for some time, until the standard is better shaped/formed. A natural part of the process is to go to SDO (Standards Developing Organizations).
Standard (Lead: Sarah Bird)
Participants: Jed Miller (Transparency Accountability Initiative), Kees Oude Lenferink (Ministry of Foreign Affairs, Netherlands), Roderick Besseling (Cordaid, Netherlands), Jeni Tennison (Open Data Institute), Ian Makgill (Open Spending)
Jed has been working on extractives and had been thinking and holding discussions last year on how to handle data interoperability. Ian is building neural networks to match company networks to companies house — 180,000 companies in the UK data set. Interested in discovering how to do unique identifiers for government departments?
There was an agreement that the way to do any generic identifier in the absence of unique identifiers:
- Freestyle field
- Prefix / System for ID (maybe a URI)
We noted that there is a difference between who the payment check went to and who the contract is with. We also need to think about historical company IDs. Cordaid has been thinking about results frameworks and the challenges with making that data comparable & aggregatable. We concluded that we need to think about how things should be formatted and structured so that people can get value.
The main takeaway points from the standard discussion are that we:
- Agree that we need a system for a typical / generic way to handle unique identifiers (in the absence of unique identifiers) — which hopefully has sufficient redundancy to link in the future.
- Agree with Jeni that a simple parent_unique_contract ID would be sufficient to enable us to do nesting of contracting processes (e.g., in the case of sub-contracting).
- Will have a meeting with Kees and Roderick to discuss results in International Aid Transparency Initiative (IATI) and what we can learn from that for open-contracting where we will also have milestones & deliverables. We also discussed at that meeting the possibilities for linking IATI and Open Contracting — where does IATI stop and contracting start.
Demand (Lead: Lindsey Marchessault)
The participants first introduced themselves and what they hoped could be achieved using contracting data. The group was comprised of government representatives, journalists, researchers, monitors and campaigners. Two strong use cases that came out of these introductions are the desire to:
- Achieve value for money
- Detect corrupt and fraudulent activity
We broke into two groups. The V4M group discussed the motivations of the government to adopt open contracting principles and standard — wanting to better track spending, more efficient procurement processes, lowering barriers to entry for small and medium sized enterprises, to get more competition and lower prices. The anti-corruption group discussed identification of collusive and corrupt practices and some of the approaches of identifying indicators in procurement data that are red flags of corruption. The input from these discussions is incorporated in the User Cases for the Open Contracting Data Standard.
Tools (Leads: Ana Brandusescu & Steve Davenport)
Participants: John Adams ( DFID, IATI), Mark Brough (PublishWhatYouFund), Chris Adams (Product Science), Eva Vozarova (Fair-Play Alliance, Slovakia)
We discussed 3 main questions:
- What tools are needed make open contracting supply-side and demand-side work?
- What are the tools that drive adoption?
- What is your experience with previous tools? Good? Bad?
The group found that the main tools to make open contracting, both supply-side and demand-side to work are: aggregators, toolkits (for frameworks), cleaning, crawling tools / scrapers and validators. For validators, it is important for us to be able to link the data between different organizations. Therefore, when we look for appropriate tools such as validators, we need to ask: Can this contract get used by another organization? We also need to handle contract datasets that have different file sizes.
The tools that drive adoption tend to be tools that provide effective visualizations (e.g., box stop visuals). From previous work experience with IATI, John talked about the lack of tools available to produce information for multiple activities. In general, there are many challenges ahead.
We need to:
- Find ways to handle inconsistencies in contract phases and data availability (e.g., no unique ID to link contract to procurement)
- Handle heterogeneous e-gov portals and open e-portals
- Competitive analysis for the private sector
- Focus on country systems
- Have multiple languages (e.g., How do we localize original language data into the standard?)
- Political pressure and peer pressure
- Consider the possibility of having a data store / API
We concluded that it is important to find the appropriate tools that will facilitate the adoption and sustainability of the standard.
To read more about our OKFest session, check out our Hackpad and Etherpad notes.