Week 3: 11-08-2017


Monday was the data and search extravaganza where the team and our boss sat down with our boss’s boss and our boss’s boss’s boss to talk through what we’ve done and what we plan to do next.

The afternoon went well and gave us a chance to show all our work from the last 16 months. Ben gave the first outing of Dia’s “101” presentation in which we attempt to explain and contextualise our work for a broad audience. Every topic from domain modelling to corporate data was covered. Tracey was particularly interested in improving internal data and costs.

Dia, Robert and the rest of the search team held a workshop to help identify user interface development candidates for the next 3 months. The search backlog is now documented and the first sprint starts Monday.

Dia also met with the business systems team to plan how the new search service will go live.


Samu and Chris added region queries to the API. This will allow users (including the new website) to find constituencies and associated MPs by NUTS level 1 regions.

Jianhan continued work investigating the current state of data around committees. Some of this information is captured as structured data but some only exists as Content Management System generated pages. And there’s an overlap. So they’ve screen scraped the 188 committees listed on the current A-Z pages and compared them with records stored in the Members Names Information Service (MNIS) and mapping data stored in the Committee Tracker database. Only 22 of the CMS pages could not be matched and are now being checked by clerks in boths houses.

People in the Commons and Lords are also checking missing links in the MNIS data which should store URLs of CMS generated pages but occasionally doesn’t. The Lords people have fixed all their missing links and the Commons people will be fixing theirs shortly.

Clerks in both houses are also checking the lay member data in MNIS.

Sarah, Aidan and Michael had a meeting with Iain Collins to discuss possible prototyping work on better interlinking between written records and parliamentary A/V, hopefully looking at forced time alignment and speaker (voice box) recognition.

Tony started work on visualising the subject indexing data (provided by our colleagues in the Indexing and Data Management Section), looking at subject cooccurrence across parliamentary outputs. His working notes are here.

Domain modelling

Anya and Michael visited GDS to try and help Ganesh (ex of this parish) domain model the process flow of government policy. Two policy experts kindly leant their brains. Many whiteboards were filled.

MySociety and Wikimedia are running a Wikifying Westminster Workshop in London on Saturday, 19th August.

ODI Cardiff are starting to run regular co-working days for people dealing with data.