It’s around this time of year that we normally publish our responsiveness statistics on WriteToThem. However, if you’ve been looking forward to seeing your MP’s ranking, we’re afraid you’ll have to wait a little longer.
Two weeks after you use WriteToThem to contact a representative, we send you an automated email to check whether or not you received a response. The data gathered by these questionnaires gives us a snapshot of how well the site is working for its users; it also allows us to highlight which MPs, which parties, and which parliamentary bodies do the best and worst at responding to constituents’ messages.
We’ve habitually analysed a calendar year of responses, January to December. Last year, though, was an election year, meaning that several MPs were active up until May, and then several new MPs took their seats in the new Parliament. So we’re going to run the data in June, looking at May 2015 to May 2016, followed by a four-week period to ensure we’ve received all the questionnaires.
Now, in theory, it shouldn’t matter too much, because we rank MPs by the percentage of mail sent through WriteToThem that they respond to (or more accurately, that our users tell us they have responded to). An MP may have responded to 100% of all their mail and then been voted out; their successor may then respond to 10% of their mail: both MPs would be ranked accordingly.
In fact, that’s how we did it for 2005, the first year for which we published WriteToThem rankings, and also an election year*.
But shifting the date like this means that the data will be less confusing. It’ll let us see how every current MP has performed, in terms of responsiveness, across a full year.
Of course, one side effect of this is that if you’re an MP and you want to be top of the pops, you have an extra five months in which to boost your score… so, on your marks, time to get writing!
*2010 fell within a four-year period during which we didn’t publish rankings.
How is the data explosion transforming our world?
That’s the question that inspires the Big Bang Data exhibition, running from today until February 28 at Somerset House in London.
Alongside all kinds of data displays, data-inspired artwork and data-based innovations, the exhibition features our very own FixMyStreet and TheyWorkForYou as examples of websites that are using data for the common good.
The exhibits range from fun to thought-provoking to visually rather beautiful: we enjoyed Nicholas Felton‘s annual reports about himself, the Dear Data project, and innovative devices such as the fitness tracker for dogs. Most of all, of course, we enjoyed seeing our very own websites put into context and available for everyone to have a go with.
We’re delighted to have been included in this event, and we recommend a visit if you’re in the area. There’s plenty to keep you interested and informed for a good hour or two.
Scottish Parliamentary proceedings are now back on TheyWorkForYou.
Back in August 2014, the Scottish Parliament changed the way it published the Official Report of its debates.
TheyWorkForYou works by fetching data from various parliamentary sources—and in this case, unfortunately, the change at the Scottish Parliament end meant that our code no longer worked. We replaced our ‘debates’ section with an apologetic note.
Well, thanks to the Scottish Parliament kindly republishing the data in almost the format we used to use, we’ve managed to make some small tweaks and restore that content—including debates from the previously missing period. If you’re subscribed to alerts, you should have received an email digest with links to the backdated content (always supposing there was any that matched your chosen keywords).
And if you’re not subscribed to alerts? Now is a great time to rectify that. We’ll send you an email every time your chosen word or phrase is mentioned in Parliament, or every time your chosen representative speaks.
While we were doing this work, we also modified TheyWorkForYou so that it now pulls in ministerial data from the Scottish Parliament API. This is a welcome time-saver for us: previously we were creating a list manually from the official PDFs, while we can now automatically fetch it and reformat it into Popolo JSON, meaning it’s consistent with all our other data.
Thanks for your patience; we know that many people were awaiting this repair, and for longer than we would have liked. Enjoy!
Ever feel sorry for the less popular kids at school?
Excellent, then you’re just the sort of person we need: you may empathise with some of the countries on Gender Balance that aren’t getting quite as much attention as the rest.
Thanks to our recent data drive, Gender Balance now contains many more countries, all waiting for you to play.
But we’ve noticed that some countries aren’t getting quite as much attention as others. Gender Balance’s ultimate aim is to provide data for researchers, and we’d hate to feel that we had patchier data for those studying the less popular places.
So, to encourage take-up, we’ve now added a ‘featured country’ spot. Accept the invitation to play the highlighted place, and you’ll receive double points, propelling you all the faster towards a coveted place on the Gender Balance leaderboard. Time to get playing!
Yesterday we told you how the data on EveryPolitician had expanded wildly in the last week. One side effect is that there are 64 new countries to play on Gender Balance.
Our gender classification game (read more about it here) runs on politician data from EveryPolitician, so by adding a whole bunch of countries, we also expanded Gender Balance’s range.
It also means that, as those countries get played, we’ll be gathering even more informative and useful data about the proportions of women to men in the world’s legislatures.
That’s all we have to say, except, 3,2,1… get playing!
Amazing—we did it!
When we decided to mark Global Legislative Openness Week with a drive to get the data for 200 countries up on EveryPolitician, in all honesty, we weren’t entirely sure it could be done.
And without the help of many people we wouldn’t have got there. But last night, we put live the data for North Korea and Sweden, making us one country over the target.
The result? There is now consistently-structured, reusable data representing the politicians in 201 countries, ready for anyone to pick up and work with. We hope you will.
That’s not to say that our job is over… far from it! There’s still plenty more to be done, as we’ll explain below.
Here’s how it happened
Getting the data for each country was a multi-step process, aided by many people. First, a suitable online source had to be located. Then, a scraper would be written: a piece of code that could visit that source and pull out the information we needed—names, districts, political parties, dates of office, etc—and put it all in the right format.
Because each country’s data had its own idiosyncrasies and formatting, we needed a different scraper for every country.
Once written, we added each scraper to EveryPolitician’s list. Crucially, scrapers aren’t just a one-off deal: ideally they’ll continue to work over time as legislatures and politicians change.
The map above shows our progress during GLOW week, from 134 countries, where we began, up to today’s count of 201.
mySociety’s Tony, Lead on the EveryPolitician project, worked non-stop this week to get as many countries as possible online. But this week we’ve seen EveryPolitician reach some kind of momentum, as it takes off as a community project. It’s an ambitious idea, and it can only succeed with the help of this kind of community effort. Thanks to everyone who helped, including (in no particular order):
Duncan Walker for writing the scraper for Uganda; Joshua Tauberer for helping with the USA data; Struan Donald for handling Ecuador, Japan, Hong Kong, Serbia and the Netherlands; Dave Whiteland, with ThaiNetizen helpfully finding the data source for Thailand; Team Popong for South Korean data; Jenna Howe for her work on El Salvador; Rubeena Mahato, Chris Maddock, Kätlin Traks, François Briatte, @confirmordeny, and @foimonkey for lots of help on finding data; Henare Degan and OpenAustralia who made the scraper for Ukraine; Matthew Somerville for covering the Falkland islands and Sweden; Liz Conlan for lots of help with Peru and American Samoa; Jaroslav Semančík who provided data for, and assistance with, Slovakia; Mathias Huter who supplied current data for Austria while Steven Hirschorn wrote a scraper for the historic data; Andy Lulham who wrote a scraper for Gibraltar; Abigail Rumsey who wrote a scraper for Sri Lanka; everyone who tweeted encouragement or retweeted our requests for help.
But there’s more
There are still 40 or so countries for which we have no data at all: you can see them here. This week has provided an enormous boost to our data, but the site’s real target is, just like the name says, to cover every politician in the world.
And once we’ve done that, there’s still the matter of both historic data, and more in-depth data for the politicians we do have. Thus far, we mostly have only the lower houses for most countries which have two — and for many countries we only have the current politicians. Going into the future we need to include much richer data on all politicians, including voting records, et cetera.
Meanwhile, our first target, to have a list of the current members of every national legislature in the world, is starting to look like it’s not so very far away. If you’d like to help us reach it, here’s how you still can.
International emergency aid charity Médecins Sans Frontiers are one of the biggest purchasers of medicine worldwide, and naturally it’s important that the drugs they buy are cost-effective. Where possible, they choose generics—white label medicines that contain the same ingredients even if they don’t carry the well-known brand names: think ‘ibuprofen’ or ‘aspirin’ rather than ‘Nurofen’ or ‘Anadin’.
But when a specific medicine is only available as a patented product from a big drugs company and with an equally big price tag attached, MSF, like everyone else, has little choice but to pay.
Curiously, this turned out to be a problem that can be solved, in part, through good web design. Here’s the story.
Obviously, drugs companies have an interest in keeping their medicines under patent. As MSF explained, patents, and in particular the practice of ‘evergreening’ them (extending their life indefinitely by making slight modifications to the medicine’s make-up), give pharmaceutical companies a monopoly on pricing, and can impede access to patients who would benefit from them.
MSF’s online project, the Patent Oppositions Database (PODB) is a resource for helping people challenge medicine patents. PODB helps groups around the world to find each other and work on cases together, and to share previous examples of art and arguments used in lawsuits which may help others in future oppositions.
The site was already up, running and functional, and the concept was sound. But it wasn’t attracting much take-up. On analysis, it became clear that this was because there was no focused experience on the site, encouraging users towards the core interactions which would power the whole concept of collaborating and sharing knowledge.
Where design came in
MSF asked us to suggest improvements that would enable groups to communicate about specific cases, and to improve the sense of community. Our solutions will add intuitive user paths that lead people to existing opposition cases and the information they need, then encourage them to join in by placing discussions and information about contributors on the page.
It’s crucial for MSF that the project reaches its full potential, and with the in-depth design changes we’ve suggested, and have now been asked to implement, we know it will.
You can read more about how we approached this project in our latest case study, over at the mySociety Services website.
Public health teams, policy-makers, councillors and NGOs often need facts about a specific area. If they’re looking for data on things like the number of smokers, the demand for hospital beds, or the birth rate, they turn to their regional Health and Wellbeing Board.
These local authority committees are required to produce a document known as a JSNA (Joint Strategic Needs Assessment) every few years. It’s a snapshot of the demographics and healthcare needs of the local population, and is used by a variety of stakeholders including policy makers and strategy groups.
Like most local authority committees, the London Borough of Hackney and the City of London Health and Wellbeing Board have previously produced their JSNA as a simple read-only PDF document. But, in the digital age, they knew that there was more they could do to make this document accessible, useful, and engaging.
That’s when they called us in — not to build the final digital version of the JSNA, but to help them understand the possibilities and ensure that they were heading down the right path.
We’ve written up the whole process in a case study, so, if you’d like to know more, read on.
Just how quickly can we hit the 200 countries mark on EveryPolitician? That’s what we’ll be finding out this week, and one thing’s for sure — we’ll get a lot further with your help.
This week is GLOW, the Global Legislative Openness Week, and we’re marking it with a concerted drive for more data.
Tony, the project lead, has consistently added one new country every day since EveryPolitician launched four months ago, and now it’s time to put a rocket behind our efforts.
The site currently contains data for 134 countries. We’ll be going flat out to see how quickly we can reach 200, and as the excitement ramps up, we hope you will help spread the word and get involved, too. Tony will carry on working as hard as he can to fill in the gaps, but we need your help to get further, faster.
What is EveryPolitician?
How can I help?
- Help us find data for more countries! We don’t currently know where to find the politician data for many countries. Here’s a list of the ones we need and here’s a page about how to contribute. If you get stuck, give us a shout.
- Write a scraper If you have the know-how, you can help us enormously by helping scrape the data from the places we do know about. See this page for guidance on how to go about writing a scraper. You’ll find lots of examples here.
- You can also help by spreading the word – tell your friends, tweet, blog, get up on a platform and talk, and just generally share this post. Thank you!
Why do we need this data?
Politician data is readily available for most countries, but it comes in a massive variety of inconsistent formats. Most of those formats aren’t ‘machine readable’, that is to say, the data can’t readily be extracted and re-used by programmers, and pretty much every country differs on what information it provides about each politician.
That being the case, anyone who wants to build an online tool that deals with politicians from more than one country, or who would like their tool to be available to people in other places, or would like to adapt an existing tool to be used elsewhere, would first have to adapt their tool to cope with the data.
EveryPolitician saves them the trouble, and the structured format also means that the tools they build will be compatible with any other tools that use it.
What kind of tools?
EveryPolitician data will be useful for all kinds of projects.
It’ll be much easier to build a website that shows people how to contact a politician. Or one that holds a government to account and educates people about what politicians are doing. Or one that helps voters make choices by displaying facts about what their politicians believe.
It can go further than that, though — with these building blocks in place, developers can really use their imagination to put together all kinds of projects, many of which we haven’t even begun to imagine. And don’t forget that, if a tool has been built to use the standardised data, it’ll also be easy for others to redeploy elsewhere.
If you’d like to see a concrete way in which the data’s already being used, check out Gender Balance.
How can I keep up to date?
We’ll be putting out regular updates via Twitter as the number of countries covered increases — plus you can watch the map turn green on http://everypolitician.org/countries.html as we progress.
Back in November 2013, we asked you what improvements you’d like to see on TheyWorkForYou.
One answer dominated: you wanted more information about how MPs vote.
Adding information on voting has been the single biggest project on the site since its launch, and has required several different phases of development. We announced each of these as it happened, but now that we’re at the end of this large piece of work, it seems like a good time for a complete overview.
So let’s take a look at exactly what it has involved—and, more importantly, what it means for you.
We’ll start with a rundown of features, then go into more detail about how they are created at the end of the post, for those who are interested.
What vote information means for you
1. You can easily see how your MP voted
Just how much do you know about how your MP voted on the stuff that matters? Most of us would have a hard time keeping up with every vote, simply because it isn’t information that’s widely publicised.
On TheyWorkForYou, you can see a run-down of how any MP has voted on key policies, by visiting their page on the site and clicking the ‘voting record’ tab (see image, above). We’ve created summaries of their stance on all kinds of matters, including the EU, same-sex marriage, NHS reform and a lot more.
Each of these summaries is compiled from every vote the MP has made on a motion that impacts on that policy.
You can click ‘show votes’ (see image above) to see the specific votes that go to make up any particular stance, and we’ve laid them all out in plain English so that it’s easy to grasp exactly what the issue is.
And from there you can click through to the website Public Whip, where you can explore votes in more detail, including lists of who voted for or against any given motion.
2. You can find out how strongly your MP feels
When we first presented voting information, we said that an MP had voted ‘strongly for’ or ‘moderately against’ certain policies, which led to quite a large postbag from people asking, “How can you vote strongly, surely you either vote for or against?”.
We wrote in the second half of this blog post about the wording changes we made to clarify the fact that these stances are calculated from a number of votes.
3. You can assess if your MP is a sheep or a lone wolf
We’ve pulled out all the votes which differ substantially from the way that the majority of each MP’s party voted. If your MP has voted against the flow, you’ll see something like this on their page:
Why do we highlight this type of vote? Because we think they’re a really good indication of where an MP feels strongly enough about something to risk sticking their neck out. It’s also a great way to check the truth when people say, “MPs? They’re all the same”.
4. You can understand the background to the votes
Generally speaking, there’s a debate before any vote takes place in Parliament, covering all the matters which may be topmost in MPs’ minds before they cast their lot.
Clicking on the ‘show full debate’ link from the topic pages (see image above) will give you the full context.
How we compile vote information
If that all seems nice and simple, well, great! That was our aim.
Putting it all together definitely wasn’t so simple, though. Voting information has never been previously presented all in one place in quite this way before—on TheyWorkForYou or anywhere else, to the best of our knowledge—so we had to figure out how to import the data and how best to display it.
As with much of our work, it’s a mixture of manual graft and automating whatever we can. Some things, like rewriting votes so that everyone can understand them, can’t be done by a computer. But many of our users are surprised to learn just how much of what we publish out is untouched by human hand.
Our Developer Struan, who did the most recent round of work on the voting records, said:
We get all our voting data from PublicWhip, a site set up by Francis Irving (once of mySociety) and Julian Todd. Public Whip takes the data we [TheyWorkForYou] produce from Hansard and extracts only the information on votes (or divisions in Parliamentary jargon) that take place in Parliament. It then allows you to look up how an MP or a Lord voted.
Let’s just think about that for a moment. We’re looking at a process where Parliament publishes Hansard, TheyWorkForYou scrapes the data and re-presents it, Public Whip extracts the voting information and presents that, and TheyWorkForYou takes that voting information back for its own voting pages. Simple…
One of the first things we did was to ‘translate’ the votes into plain English, so that it was very clear what was being voted for or against— and if you want to read more about that process, we talked about it in a blog post back in July 2014.
That allowed us to move to the next phase, as Struan explains:
Public Whip groups related votes together into policies, e.g renewing Trident, so you can see how an MP voted on the policy as a whole.
It does this by saying which way an MP would have to vote each of the divisions in the policy if they agreed with the policy. It then takes the MP’s votes on each division in the policy and assigns a score to it based on how they voted. These scores are then added up and compared to the score they would get if they always voted in agreement with the policy. The closer the MP’s score is to the score of an MP who always voted in agreement with the policy, the more they agree with the policy.
Thanks to Public Whip’s grouping, we were able to start compiling our MPs’ voting records along those same policy lines.
One of the most fiddly parts of the process was figuring out how to ensure that the information we present is a true, non-biased representation of the MP’s intentions. You might think that a vote is quite a simple matter – it’s either a yes or a no for a particular motion. But as soon as we started displaying votes within a policy, things got a bit trickier.
Some divisions in a policy can be marked as important and voting with the policy in those divisions is worth more points. This is to prevent voting in agreement on a set of minor votes, e.g “Parliament will commission a report on the future of Trident”, outweighing voting against something important, e.g. “Renew Trident”. It also reflects the way Parliament works, often with several smaller votes on parts of a bill and then a vote on the bill as a whole.
For clarity I should point out here that sometimes voting no in a division is a vote for the policy, e.g voting no in a “This house believes Trident should not be renewed” division would clearly be a vote for our example “Renew Trident” policy.
This approach also helps where one vote straddles several topics: for example, consider a vote against the Budget when the Budget contains many proposals including, say, the capping of VAT. It’s quite possible that an MP may be for the capping of VAT but broadly against several other motions covered by the Budget, and so decides to vote against it on balance. So long as we mark the Budget vote as a weak vote for the capping of VAT, its significance should be properly accounted for.
Where we don’t have enough information to show a stance, for example where an MP never voted on the topic, is too new to have had a chance to vote on the topic, or all their votes on the topic have been labelled as ‘weak’, we say so:
A final little subtlety is the difference between “Never voted” on a policy and votes where the MP was absent. If it says an MP has never voted on a policy that means they were elected after all the divisions in the policy took place so did not have a chance to vote on them. Absent means they could have voted in the divisions but did not.
Absent votes count towards your score but at half the rate of voting in agreement with the policy. This is so that an MP who votes in agreement with the policy in one division and then misses all the other divisions shows as agreeing with the policy rather than against as it would if no score was assigned to absent votes. That does currently mean that if they were always absent it shows, slightly unhelpfully, as “a mixture of for and against”.
It’s not an ideal system as it does produce some odd results occasionally but it mostly works.
To show where an MP has voted against the majority of their party, we have to figure out a similar score across the party as a whole.
This is exactly the same process, only we add up all the votes by all the MPs but the maths is pretty much the same.
All in a day’s work
As mentioned at the top of this post, vote information was our most-requested addition. And rightly so! Our MPs represent us, so naturally we want to see their track records, quickly and easily.
If you’re not an expert, you might not have known how to find this information before. And that’s essentially what TheyWorkForYou aims to do: make the workings of Parliament more accessible for everyone.
Parliamentary copyright images are reproduced with the permission of Parliament.