South Africa’s Parliament produces a significant volume of information. Committee minutes, attendance registers, written questions to ministers, budget review reports and so much more. This written record of parliamentary activity is extensive, and the problem is that it isn’t easily accessible. It’s information, not data, and the latter is critical to interpreting the results and moving accountability from opinion to evidence.
“Parliament is good at sharing information, their website is rich with it, but the problem is that it is very qualitative,” says Dirk Meerkotter, Developer at OpenUp. “It is information locked inside PDFs or Word documents and it is hard to extract actual metrics from these documents. Without data, criticism of Parliament is just assertion, but with data, it is possible to prove accountability and to provide a baseline that allows for citizens and institutions alike to measure change in either direction.”
The information locked inside these unstructured documents can’t be searched, compared or tracked over time. This makes it challenging to detect trends or patterns of poor performance, and this breaks the pipeline at every stage from capture to structure to publication. At capture, data that should be structured from the moment it is recorded, arrives in inconsistent formats and is impossible to aggregate. There is no common schema or standardisation or way of connecting one dataset to another, so structure is missing. And at publication, what reaches the public is a document rather than data, which means it can be read and filed but not interrogated.
For example, individual voting records, which are standard practice in parliamentary democracies around the world, don’t appear to exist at the member level in local Parliament. Parties report aggregate yeas and nays, making individual accountability technically impossible, while members’ interest declarations are circulated as Word documents for manual entry with predictably inconsistent results.
“The word ‘none’, when indicating something is not applicable, is spelled five different ways,” says Meerkotter. “And it is done this way because this is how it has always been done. What makes this particularly striking is that structured data infrastructure does exist within Parliament’s systems, it’s just not publicly accessible. APIs run in the background of certain parliamentary platforms, including systems for tracking written questions, but they were never intended for public use.”
This is the gap that ParliMeter was designed to close. The organisation was developed in partnership with OUTA, OpenUp and the Parliamentary Monitoring Group (PMG), and co-funded by the European Union. Its purpose is to take years of accumulated parliamentary records and convert them into searchable, visual data. Thus turning what was previously an archive of frozen documents into a living picture of how South Africa’s MPs, parties and portfolio committees are performing. And for the first time, it is possible to answer questions about Parliament with evidence.
ParliMeter’s primary source of data is PMG, which attends every committee meeting, records attendance, takes notes and writes summaries. For questions data, PMG records are combined with Parliament’s own monthly register – an e-mail dispatch flagging questions answered on time, answered late or lapsed entirely – to build a picture of ministerial responsiveness that neither source can provide alone. Working with this data requires making interpretive choices and Meerkotter is candid about the implications: “Whenever you work with data, you make decisions about how you clean it and with that comes the possibility that someone else could do it differently. Take attendance, for example; a 50% attendance rate means something very different for an ANC MP than it does for an MP from a smaller party.”
This is because bigger parties like the ANC can ensure they have multiple MPs in a committee, while smaller parties cannot have someone in each committee and a single MP might be a member of multiple committees. If that MP attends 50% of the meetings of one committee, it might be because it’s physically impossible for them to attend more than one meeting at the same time. So an MP from a small party might actually have 100% attendance of the meetings that they’re able to attend, yet it will still show up as 50% attendance.
For now, ParliMeter is able to track attendance, as well as provide indicators to attempt performance measurement via the Bill Tracker, the Q&As dashboard and the Indicators for Democratic Parliament (IDP) reports and assessments. This is a starting point but falls well short of what meaningful oversight requires. The more valuable indicators, individual voting records, the quality and content of questions raised in committees, members' interests and financial disclosures, remain either uncaptured or effectively locked away.
"There is not enough external pressure to change," Meerkotter says. "There is no obligation to do so, and Parliament can always claim transparency because the information is on the website, but it is not usable."
South African citizens are currently making their most fundamental democratic decisions such as who to vote for – whether their representative is doing their job, whether the party they supported is keeping its promises – without the right evidence. And this is why ParliMeter exists. It fills the gap that ensures information becomes measurable and meaningful.
Share
