Originally posted on Civicist. This post is a summary of the research we presented at The Impacts of Civic Tech Conference 2019 at the OECD in Paris, France. That presentation video and interactive timeline are available here. We also shared this work at Code for America Summit in Oakland this week. This work and its outputs are CC-BY licensed to Matt Stempeck, Micah Sifry, Aliya Bhatia, and Sruti Modekurty.

Since re-launching the Civic Tech Field Guide at Code for All Summit last fall, we’ve been working on an ambitious, data-driven timeline rooted in a collective look at our entire, broadly-defined field. This work builds on our earlier research into the Civic Tech Graveyard and issues with impact measurement in civic tech.Got a correction? Wonderful – just edit that project directly in the Civic Tech Field Guide, or let us know.

This is our timeline of civic tech, spanning 25 years from 1994-2018. It tracks discrete product and project launches, not just organizations. The data consists of categorized project launch dates, not end dates, so we’re not exploring project longevity here.We’ve broadly divided the collection up between the tech (products, apps, websites) and the social (everything that people do: the conferences, funders, meetups of civic tech). We also track adjacent and foundational projects that aren’t explicitly civic tech, but are necessary for civic tech to have meaning: issues like broadly available internet connectivity and racial, gender, and linguistic diversity online.

 

This timeline is rooted in the data available in the Civic Tech Field Guide, a crowdsourced, open global collection of more than 2,000 civic tech tools and projects. Thousands of civic tech practitioners from more than 100 countries have used and contributed to this living resource. We catalog not only the tools, but also the social side of our field: the conferences, funders, awards, design principles and playbooks. One of the reasons we are building and curating the guide is to mine it for research like this.

We’re asking you to adopt your project, as well as add missing projects. Even though we’ve spent hundreds of hours on this for the past few years, there are still important projects missing. This is a living repository of knowledge.

Why create a civic tech timeline in the first place?

Micah once heard someone say, “Fish need to be their own oceanographers.” While it’s not clear what that means exactly, it is true that if you want to know where you are going, it helps to have a map. We think it’s important to begin to tell the story of our field. With so much at stake, and limited resources with which to take on major challenges, we believe it’s important to track our collective progress (and our mis-steps).

As a first step towards further understanding, we’ve done something very simple: collect the start dates of all the entities in the Civic Tech Field Guide. With this data, we are posing a few research questions. First, has the field of civic tech grown, as defined by project count? And is it still growing? Second, has the field grown more complicated, as defined by category count? And, finally, is our data useful?

To answer these questions we pursued multiple types of analysis: first, conversations with ourselves and the hypotheses we set. We took this dataset and used pivot tables to analyze the distribution of categories over the years.

Standardizing the date-founded data to only the year projects were founded, we’re able to identify the median date-founded within categories to see inflection points. These occur when the number of projects in a given category crosses the 50% mark of all organizations within that category, followed by a decline in that category’s new-project growth rate. This inflection point was then compared to the mode, i.e. when most of the organizations of a certain type were created. This foundational standardization helped us compare across categories, across time, and across more abstract aspects of civic tech, like tool development versus field development.

We know our data isn’t perfect. Caveats, we’ve got a few. Much of this data is crowdsourced data. It’s artisanal and hand-picked. There’s unevenness across categories. For example, if we were to import a dataset we recently found containing lots of Canadian innovation labs, we know it will skew our findings in that direction. Still, with about 2000 entity start dates collected, we’ve compiled the most comprehensive chronology of civic tech available.

This dataset, like everything on the Field Guide, is Creative Commons licensed. We’ve made the data available as a CSV here. If you want more complete profile data for each project, you can access it using the core WordPress JSON API. We ask that you attribute the Civic Tech Field Guide with a link, and use the data for non-commercial projects (Creative Commons BY NC SA 4.0).

Previous timelines

 

We wouldn’t be good chroniclers if we didn’t acknowledge that this isn’t the first civic tech timeline. Micah and collaborators compiled a rich timeline of important events at the intersection of the internet and politics that spans from 1968 to 2013 (image above). It’s still online at http://techpresident.com/timeline, and still well worth a look.

 

Another more recent Civic Tech and Design Timeline was created by civic designer Cyd Harrell. It looks at key events, primarily in the United States, in the evolution of civic design practice (image above).

We love these timelines for telling the story of our field. It’s important to note that any timeline built around a limited set of landmark moments will be culturally subjective.

We’ve also called out some key moments in the evolution of the field to provide historical context, where we believe they’re representative. But with an eye towards a more inclusive, more data-driven story of our field, we sought to create a broad, participatory history of civic tech.

The 1990s

We start our timeline in the 1990s, but must note that there are precursors to civic tech that predate the 90s. There was a rich conversation around Appropriate Technology in the late 1960s. The field of Information and Communication Technology for Development (ICT4D) and Development Communication before it paved the way for much of the work we do in civic tech. For our purposes, we focus on all of the civic action people can take when the internet emerges as something that many people can not only read, but write on. (We’re continuing to collect other “predecessors” to civic tech — please share your favorites in the comments to this piece!)

 

That kind of civic action starts to coalesce around 1994. This is not a coincidence: 1994 is when Mosaic, the first popular browser, was launched, and more people outside of academia began using the World Wide Web. Some key civic tech moments from the 1990s:

• 1993: Senator Ted Kennedy has the first website built for a federally elected representative
• 1996: Baltimore launches its 311 system, a precursor to many examples of cities collecting data
• 1997: e-Estonia launches, representing many examples of e-government and e-democracy in the 1990s

 

One dominant pattern that shaped the emergent field of civic tech: early pioneers try something novel that strikes a chord, and it soon grows into an entire genre of civic tech practice.

In 1993, when the Securities and Exchange Commission dragged its feet in making its information available, Carl Malamud bought the dataset and published it as a freely-available resource online. When his small NSF grant ran out two years later, he told his active user community, if you want this to continue, contact then-Vice President Al Gore and the SEC and tell them this should be a government service. After a public campaign and a lot of back-and-forth, the SEC took it over.

Twenty five years is a long time in technology. That’s several generations. While investigating the data, we found a few interesting examples that prove that point. For example, several different ventures have launched and shut down at a single URL: evote.com. And over the years, the people-powered Code for America Brigade has won the namespace against the tens-of-millions-in-funding Brigade.

The 2000s

Heading into the 2000s, we survived Y2K and now civic tech enters into a rich era of technical experimentation. This is the golden age of map and app mashups, the introduction of the major social media platforms, and people throwing spaghetti at the wall, or “making nifty code and putting it online for others to use,” as mySociety’s Rebecca Rumbul put it in her 2019 TICTeC keynote.

We see more important examples following the pattern mentioned earlier in this piece, where innovative experimentation becomes common practice and helps define the field:

• In 2004, the predecessor of TheyWorkForYou spawns other mySociety projects that get adopted around the world, and used by millions of people.
• In 2005, two civic map mashups, Housingmaps.org and Chicago Crime Map, predate the Google Maps API. With the free API (until 2018) and the development of OpenStreetMap, we see a boom in civic maps.
• In 2006, the Sunlight Foundation is founded, launching with a commitment to opening up government data, starting with two key projects, OpenCongress and Congresspedia, that were inspired by mySociety efforts.
• In the winter of 2007-2008, as the Kenyan elections teetered into chaos and the government clamped down on reporting, Ory Okolloh and several collaborators created Ushahidi. While there were other examples of people coordinating online to organize digital participatory aid responses to crises using tools like blogs, it was Ushahidi that turned that practice into a platform solution, supporting not just crisis response, but election monitoring, municipal issue reporting, and more.
• In the fall of 2008, Vivek Kundra, then CTO of Washington, DC, hosted the first open government data hackathon. Apps for Democracy invited outside civic hackers to make use of the city’s open data portal, leading to the creation of 47 apps. The event likely inspired over 50 other such events around the world.
• In the middle of 2009, Jen Pahlka created a fellowship program to bring tech talent into serving government with Code for America, leading to a renewed emphasis on improving government services from within institutions.
• Towards the end of the decade, we see BarCamps and more social projects. Code for America launches in 2009 to hack government bureaucracy in addition to hacking code.

The 2010s

This is the decade of the civic tech project boom, which actually starts in 2008 and accelerates in the early 2010s. The median start-date for social civic tech meetings like Meetups is 2013. The year 2015 sees the most projects launched of any year in the database. Even tech giants get in the game, like Google Civics, Facebook’s Civic Engagement team, Mozilla Advocacy, and Microsoft Cities.

The 2010s is also the decade where civic tech becomes self-referential, Beginning in 2014 and 2015 we see lots more references to ‘civic tech’ (as well as debates about what that phrase means). The Big Bang Moment for this development is probably the December 2013 publication of an oft-cited Knight Foundation report, analyzing the emergence of civic tech and where money is being invested in the space. We see more alignment in terminology, with more groups explicitly using ‘civic’ in their organization names.

The mid-2010s is also when the knowledge development side of our field begin to explode. The guide includes media about the field of civic tech, including the launches of books, courses, newsletters, blogs, and podcasts about civic tech. 2013 is a big year for the number of launches of projects centered in this knowledge creation. Aliya posits that this spate of cultural production is an important predecessor to the technological development that follows.

2013 is also an important year for the launch of civic hacking meetups, many of which are also more explicitly branded as civic tech outfits. The social side of the field really took off in the last 5-6 years. In 2015, Civic Hall launches as a community center, and Apolitical launches as a government innovation resource and peer network.

To return to our initial research questions, first, clear the field has grown.

Looking at ten years of the Tech and the Social, we see a boom of new projects launched starting in 2007. Here’s a histogram of the crescendo of new projects launched per year (not total projects in the field):

In our dataset, the year with the most project launches is 2016.

Second, is the civic tech field still growing?

The answer is yes, but also that its rate of growth may be slowing. It peaked in 2018, after which there is a bit of a decline in new projects launched, which we’ll explore.

In this graph, the green histogram bars are the number of absolute project launches per year. The blue line is the trend, or the year-over-year growth compared to the previous year. That’s where we see, in 2017 and 2018, a bit of slowdown in the rate of new projects launched.

So what explains the apparent slowdown? An optimistic interpretation could be that this slowdown is occurring because the field (and its funders) are beginning to understand a bit more about what works, and fund those projects rather than create as many new experiments. It’s also possible that there are fewer experiments, because there existing projects are taking up that space. Or, that there are certain parts of the wall where the spaghetti just won’t stick, like non-partisan political social networks, and people have finally realized it. Our data isn’t comprehensive enough to prove which, if any, of these hypotheses might be correct.

It is also entirely possible that we just haven’t discovered and logged as many recent projects into the Field Guide. But our original fear was the opposite: we thought we might have a recency bias because our work on data collection has been most intensive in the last year. So we actually expected 2018 to be overrepresented, not under. We’ll be able to revisit this hypothesis a year from now, when more time has passed since 2018.

Clearly the field has grown more complicated, as measured by the number of categories of civic tech projects in existence.A field born in open government data portals and civic apps has matured.

Here’s the table of contents on the Civic Tech Field Guide:

This may look like an overwhelming number of subcategories, but in our minds it’s worth separating out work being done to transparently visualize government budgets from electoral organizing campaigns using peer to peer SMS to elect a candidate, even though both examples are employing tech for democratic purposes.

In 2008, civic tech projects were launched across 43 different subcategories of civic tech. By 2018, projects were launched across 72 various subcategories.

We’re seeing increased sophistication in the field in a number of different ways. Entire NGOs have been spun up to serve as intermediaries for civic tech projects. Civic tech increasingly surfaces as a method of doing the work, such as in private public partnerships employing data for the public good, rather than just discrete civic startups. Even the proverbial civic tech hackathon has become more focused. In 2014, we see a record number of civic tech hackathons, convenings, and events take place, but more recently practitioners are creating specific intentional spaces around specific topics, like the Make the Breast Pump Not Suck hackathon, as well as locations, like Seattle Tech 4 Housing.

This streamgraph shows the evolution of the field of civic tech from 1994-2018, based on the start-dates of 2,000 individual projects collected in the Field Guide as of March 1, 2019. The projects are sorted by the Guide’s 200+ categories, which you can see by mousing over any part of the streamgraph. Categories are sorted alphabetically from A to Z, starting at the bottom of the graph and moving upwards. The wider any color in the graph, the greater the number of projects that started in that category in the year below it. At civictech.guide/timeline, you can hover over each category in this rich tapestry to explore it over the years. Huge credit to Sruti Modekurty, who built this data visualization using the Javascript D3 library.

Looking forward

This dataset, like the Civic Tech Field Guide itself, is part of an ongoing effort. Right now, we have to add lots of caveats when we look to this data for insights. The data is skewed toward the anglophone world, although we now have data from many countries.

Standardizing the data to year-founded (rather than a higher, month- and day-founded resolution) was crucial, but it came at the expense of a lot of chronological nuance. We only had month founded data for a fraction of the data. We also had to consider whether to include adjacent actors like funders and public institutions who support civic tech, and whether their start dates should be categorized by the year they were founded, or the year in which they started funding civic tech. For example, the New York Public Library was founded in 1895, which could skew the entire data set.

Other limitations include categorization. We used the first category for entries that were classified with multiple categories (these are a small but growing subset of the overall collection). This means the full complexity and diversity of projects may not be represented in our analysis.

Lastly, new experiments in civic tech are increasingly high-tech. In the Civic Tech Field Guide, we also track the growing number of examples of emergent tech being used for democracy. We’re seeing more projects involving drones (6 projects) and Virtual and Augmented Reality (8 entries from 2012-2018). Cheaper sensors have allowed a flourishing or environmental data projects, like HeatSeek NYC. Many projects are trying to apply artificial intelligence toward civic goals, leading Marci Harris to point out a tension concerning how automated we want to make our consensus-based systems and institutions. And blockchain is increasingly being applied toward civic ends, like the Bitfury land registry in Georgia, Civil’s attempts to use blockchain for journalism, and several blockchain voting concepts. That said, the “Impacts of Blockchain Tech Conference” (TIBTeC) has not yet been founded.

To enhance this dataset, we’re planning to go back into the projects and identify the years in which they ended. This will give us a better sense of project longevity. As part of this work, we also need to separate one-off projects, like reports releases, from ongoing projects, like an app serving users.

Even with these limitations of our current data, we’re pleased with what we’ve been able to accomplish here. As noted, we expected to see a recency bias in the data, but we don’t. We saw fewer projects in 2017 and 2018, even though that’s the time period when we were most actively chronicling projects.

Project launches are only one way to survey the field. There are so many opportunities for future research. We also want to look at other ways of slicing the data (focusing on certain categories or geographies, for example). We’re keenly interested in collecting the data on diversity in the field, like how many organizations were founded by people of color and women, and the different effects private and public funding have in the field, just for starters.

We’d love to hear your thoughts in the comments below. You can explore the interactive version of the timeline at https://civictech.guide/timeline/. And we’d like to thank our funders, Knight Foundation and Luminate, for making this type of field-wide research possible.