Blog posts by Matthew Somerville

Matthew Somerville, Head of Development at SocietyWorks, shares his experience of speaking at a conference dedicated to discontinued civic tech and what can be learnt from it.

Read more posts from the team talking openly about what they’re up to.

___

A few weeks ago, I gave a short talk at the second Workshop of Discontinued Civic Tech, held online and in person in Japan. The topic was “What does ‘Failure’ Mean in Civic Tech?” (or should we call that pro-democracy tech?).

My talk was about PledgeBank, but I’ll get to that in a minute.

Matt Stempeck first gave a talk about the Civic Tech Graveyard, what themes and lessons can be drawn from the entries there, and how more research is needed into the current state of affairs. It was interesting to see that the top category in the graveyard was collective action, and in there that the number one example shown was PledgeBank – as that was the subject of my presentation that followed straight after.

For those who don’t know, PledgeBank was a website run by mySociety from 2005 to 2015; the second service created after WriteToThem. Its core idea was to help people overcome the collective action dilemma, of wanting to do something but being unable to do it alone; using the internet to gather people together in support. It let people set up pledges in the form: ‘I will do something, but only if a certain number of people will help me’.

With the initial launch and for some time after, plenty of work was done on the site with innovative features like SMS signing, PDF poster generation and local geolocation alerts – remember this was nearly 20 years ago now! But mySociety was and is a small organisation, reliant on a combination of funding, donations, and commercial services, and in 2015 it was decided to close some of our original services, including PledgeBank, to concentrate more on a few core services and our international partnerships at the time.

Certainly, PledgeBank did have a number of individual successes in the decade it was around. As well as various charity collection drives (such as underwear for orphans in Liberia or books to create a town library in India), notable lasting legacies of the site include the foundation of the Open Rights Group charity in the United Kingdom, and the fact that 1,000 people in the United States pledged to move house to New Hampshire as part of the Free State Project. And football fans raised over £20,000 for Ebbsfleet United, so that they could buy striker Michael Gash.

We also did some work with Barnet Council in London for a special custom version, where people could for example ask the council to approve a road closure for a street party, if enough residents of the road agreed. So it definitely proved that a website could solve some coordination problems by using the internet. And I’m unaware of any other successful socially focussed version of the same model.

But listening to all that, you’re probably thinking of some rather more ‘successful’ organisations using a similar model since then, for example Groupon, or Kickstarter.

So what happened with PledgeBank? In my opinion:

• The concept was too diffuse, and didn’t provide a reason for people to Do a Thing – pledges with a long lasting effect were more likely ones where people got something out.
• Lacked single-minded vision and the resource to implement it; we were and are a small organisation spread thinly
• Related to that, we did not do enough on the marketing side, both external and internal in terms of curating good pledges, working with potential pledgers, involving more people.
• We weren’t going to drop everything and try and turn PledgeBank into a profitable business in order to fund our charitable work, even if that might have possibly worked out in the end.

After my talk came a number of other interesting ones, including one by Gurden Batra from Dark Matter Labs (an organisation we have worked with on Neighbourhood Warmth, which has some PledgeBank related activity, where it was interesting to see them wrestling with the same issues of funding and longevity that we do.

Matt had also mentioned his Civic Tech Field Guide research – we talk all about the impacts of civic tech at our TICTeC conference, which will be held next year on 10 & 11 June in Mechelen (Belgium) and online. The Call for Proposals is now open, and hope to see you there.

___

Image: Photo taken at the workshop by Discontinued Civic Tech

At SocietyWorks we believe in transparency. One of the ways we live this value is by working in the open, and giving our team members space on our blog to write about what they’re working on, something they’re interested in or even perhaps a mistake or challenge they’ve learned from.

In this blog post our Head of Development, Matthew Somerville, writes about his experience attending the Open Data Camp 9 unconference in Manchester on 6 and 7 July 2024.

—

Last weekend, I went to the Open Data Camp 9 unconference in Manchester. I hadn’t been to an Open Data Camp before; it was very well organised, with good food, lots of volunteers, a creche, people from Drawnalism making lovely pictorial summaries of many of the sessions (see the website link above, and I’ll embed some from the sessions I went to below so you can see how amazing they are), they organised accommodation (if you needed it), and more.

For those who might not know what is meant by “open data”, there was a session about that – there’s a really good summary in the session notes at Open Data 101: Open Day for Newbies (2024 edition). The definition given there is: “Open data is data that can be freely used, reused and redistributed by anyone, subject at most only to the requirement to attribute and share alike.”

Here at mySociety and SocietyWorks, we use, reuse, and publish a lot of open data. For example, MapIt is based on open data, as is FixMyStreet and FixMyStreet Pro. TheyWorkForYou is repurposing open data into a slightly better format and WhatDoTheyKnow also includes a lot of open data.

The venue was the Engineering building of the University of Manchester, which was round the corner from where I went to school (more on that later), and perfectly designed for an unconference, with four separate rooms all coming off a central hub room for food/ drinks/ chats. They had a Lego board to show where people had come from, and a pile of old out-of-copyright Manchester maps.

At this unconference, the pitches were ideas that people wanted to talk about and discuss as a group with interested others – I was happy just to see what came up and hopefully have some interesting conversations.

Day one

In the morning, I first went to a talk about deleting data and having too much data, which was a broad look at the costs of maintenance and APIs vs datasets. I raised the idea of it being much easier to maintain/look after if the open data is embedded within the processes of that data (e.g. your street light asset management system leading directly to the publication of that street light data, not requiring a special export to a special open data platform that could be subject to the vagaries of the current postholders). Following this I attended a discussion about the digital/data priorities of the first 100 days of the new Labour government.

In the afternoon, I went to a session about data on elected officials / elections by Open Data Manchester, who had made e.g. a poster of deprivation vs representation, and were looking at doing more with councillor information and data. I contributed some info on how TheyWorkForYou and WriteToThem works, our combined IMD dataset, and the popolo standard for representative data.

Then it was over to Owen Boswarva’s session on the campaign/case for open addresses. This has always been a topic dear and core to us; WriteToThem and TheyWorkForYou cannot provide accurate answers for every single postcode due to the lack of open address data. I/we were well-known by everyone there, and it was a look at the current situation and what could be done to push this forward. The new government is of course one possibility, and the new Business & Trade Minister (in charge of Royal Mail, if not Ordnance Survey) has met with people on this exact topic.

The last session I went to on the first day was about web scraping, open data, and ethics. Lot of self-awareness at this, looking at my and our history with TheyWorkForYou, Mapumental, traintimes, Theatricalia, someone else’s project on scraping Warm Spaces locations, and what differences are there in terms of ethicalness and behaviour.

Day two

Day two, after catching the same bus I used to catch as a kid to school (ever so slightly more expensive now), I went to a session by two people from Raileasy, wanting to talk about open data success stories in public transport. Lots of good chat about train data, bus data and the pros and cons of decentralisation.

Being that it’s a phrase we use often here at SocietyWorks when talking about what we help local authorities with, I couldn’t not go to a session called “Closing the feedback loop” by someone from Open Data Scotland discussing how do/can producers of open data be made aware of how their data is used; e.g. in the government case, generally so they know they shouldn’t just turn it off (though turning it off does bring people out of the woodwork, certainly!). Other possibilities discussed included asking for an email as ‘payment’ for getting the data, and in order to get notified of updates or deletions; or having a place to show/link to examples of how that specific data is used.

After lunch, the organisers ran a “go outside and explore” session to try and notice things you might not normally notice, with an animal avatar. I wanted to go back and see my old school, so I co-opted the octopus group to do this, and we had a nice walk around the area (which again, is quite changed from the 1990s and the Crescents), finding a wildflower meadow while we discussed open data.

Lastly, I went to data horror and data joy stories, where you can probably imagine some of the things talked about – one thing I mentioned was the opening up of Bank Holiday data in an official GOV.UK JSON file, which meant I could submit a Pull Request on GitHub when there was a mistake, and from there find out that Scotland had forgotten to create a Bank Holiday in 2010 and 2011…

Other sessions I didn’t go to, but would have liked to…

• Refreshing our approach to government transparency
• Sustainability: how can open data help?
• Public/private sector marriage guidance (from, understandably, a data point of view)
• AI: Ethics and implications for open data
• Open data, communities, and the role of gamification

And probably more – do take a quick look through their blog.

That’s it! Thanks for having me, Open Data Camp!

At SocietyWorks we believe in transparency. One of the ways we live this value is by working in the open, and giving our team members space on our blog to write about what they’re working on, something they’re interested in or even perhaps a mistake or challenge they’ve learned from.

In this blog post our Head of Development, Matthew Somerville, writes about a new waste container generator for our WasteWorks solution which uses CSS to make it easier to generate waste images tailored to individual councils’ branding and bin types.

—

WasteWorks is in use by a number of different councils, all of which have their own types of bin, in various colours and sizes (you can see a large array of photos of bins in use by local councils at the lovely govbins.uk site). An image of each type of bin associated with a property is displayed to residents from the WasteWorks homepage.

Our designer Lucas came up with some nice simple iconography for us to use on bin day pages, including domestic wheelie bins, communal bins, sacks and boxes. When we’ve had a new client, whose bins are differently coloured to any previous council, Lucas has been providing us with new PNG images to match the bins the council uses, exported from the source vector images. As well as the right colours, each PNG had to be provided at two different sizes, to work with high-resolution displays, and so we’ve built up a small collection of such bin images over the years.

These pictures only vary by colour (and the presence of a recycling logo), so I wondered if there was a better way we could generate these images. SVGs are vector graphics – they scale to any resolution, and importantly for this can be styled with CSS, the same mechanism used to style a web page. CSS also has “variables”, where you can define e.g. a variable to be a particular colour, and then use that variable in a different part of the document (or SVG image).

Lucas and I worked together to come up with SVGs for the various containers, which instead of specifying any fill colours directly, used one or more CSS variables to specify the colours. For the recycling logo, we set a variable for the opacity of the logo – defaulting to 0, so invisible, but we can set it to 1 to have it appear.

We’ve added a page to our user manual where you can play around with the colour of our new bin images in real time:
https://www.societyworks.org/manuals/wasteworks/container-picture-generator/ :-)

More technical details

The header of our domestic wheelie bin SVG looks something like this:

<svg class="waste-service-image">
 <style>
  .wheel { fill: #333333; }
  .main { fill: var(--primary-color, var(--default-color)); }
  .lid { fill: var(--lid-color, var(--primary-color, var(--default-color))); }
  .recycling-logo { fill: #ffffff; opacity: var(--recycling-logo, 0); }
 </style>
 ...

See that the CSS uses the var() fallback parameter so that the lid colour will e.g. default to the primary colour if not specified.

We used CSS mix-blend-modes in order to have shadows and highlights that would work regardless of the colour they were placed on top of; here’s an example shadow:

<g style="mix-blend-mode:multiply" opacity="0.12">
  <path fill="black" d="M95.1227 90.7581L120.495 681.301H79.3654L50.9626 90.7581H95.1227Z">
</g>

Then when we embed an SVG in someone’s bin day page, the web page itself, outside the image, can specify what colour to use by setting the corresponding CSS variable, and the picture will then appear in the right colours. When we have a new bin colour, we don’t need to create a new image, only set the right colour.

Here’s the outline of an SVG on a bin day page, for a grey bin with a blue lid, showing the recycling logo:

<span style="--primary-color: #767472; --lid-color: #00A6D2; --recycling-logo: 1;">
 <svg class="waste-service-image">
  <style>
   [... style as above ...]
  </style>
  [...]
 </svg>
</span>