In his book, How Innovation Works, Matt Ridley suggests that innovation has been critical to advancing human prosperity over the past few hundred years. I largely agree with this perspective, and I would add two additional points:
- Innovation is about more than technology; it encompasses governance, contracts, processes, organizational structure, markets, and capital; and
- Innovation rarely happens in moments of individual inspiration, but, rather, happens incrementally, over time, and often via fortuitous events that happen through interactions between groups of people.
Innovation is a key ingredient of Azavea’s growth and long-term resilience over the past two decades. Azavea’s mission includes two components: 1) we seek to advance the state of the art in geospatial technology; and 2) we apply that technology for positive impact.
We pursue the innovation component of our mission in a number of ways:
- Seeking out client projects that push us to develop new geospatial data analysis capabilities
- Pursuing federal and philanthropic R&D funding
- Supporting fellowship programs that result in new technology
- Partnering with academic and nonprofit researchers
- Developing long-term investments in new libraries, like our Raster Vision, Groundwork, and GeoTrellis projects.
- Cultivating an environment that enables our colleagues to create new things
In this article, I’d like to expand on this last point by outlining our 10% time program in some more detail, highlighting some notable examples and themes, and describing some of the impact the program has had on Azavea’s work.
The basics: what is 10% time?
Learning and trying out new ideas and technologies have always been key motivators for me personally. Azavea’s 10% time program, however, (like most good ideas at Azavea) was not one of my ideas. It began in 2003 with the goal of creating space for colleagues to try out new ideas and technology that they felt were important or might become important in the future. That goal has remained fundamentally the same in the years since then; the 10% time program is an invitation to reserve time to test new ideas and learn from the process.
We view 10% time as a benefit you earn by mastering the other basic skills required to do the work. In someone’s first six months at Azavea, most people have a lot to learn in general – clients, projects, code, processes, norms, etc. – it’s a lot to absorb. After the initial six months, we expect folks to continue learning and growing in a proactive and organized way, and the 10% time program is extended to all colleagues in the organization (not just engineers) after that initial 6 months. It is important to me that this initiative is not just an engineering program. I believe we need to be pursuing innovation at every level: management, marketing, finance, and data analysis as well as technology innovation. All of us need to be spending time getting better at our work.
How 10% time has played out at Azavea
Some 10% time projects have covered a range of topics and types of engagements, and some have completely changed the course of Azavea’s history. I see our 10% time projects falling into seven general themes:
- Quality of life improvements
- Exploring new tools and technology
- Creating new capabilities
- Small projects that get bigger
- Contributing to open source projects
- Pro bono projects
- Learning projects
Quality of life improvements
Working with smart and passionate colleagues sometimes has benefits outside of true work-work. Dedicated time spent on quality of life improvements have made life at Azavea a little bit sweeter. Jeff’s famous timesheet enhancements have made our weekly time entry ritual a little easier, a little faster, and a little less likely to have errors. And, to ensure that brunch lovers knew when the specials were on, Chip created a menu checker and notification for the Cafe Lift menu.
Exploring New Tools and Technology
With ubiquitous maps and directions on our phones today, it’s easy to forget that when the first iPhone was released in 2007, it did not have a GPS (or even a compass). The GPS showed up in the iPhone 3G, but things got really interesting when the iPhone 4 was released in 2010. This phone added a gyroscope and an accelerometer. Now the phone could detect its location in three dimensions as well as its orientation and which direction the camera was pointed. This unlocked some interesting capabilities, and my colleague, Josh, used his 10% time to experiment with the idea of overlaying geometric objects in the camera view on the phone. Josh initially experimented with using the MIT ARToolkit and NYARToolkit. Azavea leveraged this initial prototype, a partnership with the City of Philadelphia Archives, and funding from the National Endowment for the Humanities (NEH) to develop a mobile app that would display historic photos overlaid on the phone’s camera. When we finished, you could see these historic photos overlaid on top of the camera view, and it was pretty amazing.
In a related effort, we bought developer versions of Google Glass for Mike and Kenny, who demonstrated how both historic photos and other city data could be overlaid on our visual experience as we walked through a city.
Creating New Capabilities
A number of 10% time projects have been attempts to create new tools or libraries that make possible new capabilities. A few recent examples include:
- PySTAC: a core library for working with SpatioTemporal Asset Catalogs (STAC)
- Franklin – an open source library for serving STAC catalogs
- Loam: a wrapper for running GDAL in the browser using gdal-js
- Grout: a flexible data schema framework for geospatial apps
- TileGarden: serverless tile rendering with AWS Lambda
- PyCurb: a library for working with street parking data using the CurbLR data standard.
- Showtime: makes it easy to apply CSS transitions to React elements
Small projects that get bigger
The O.G. Azavea 10% time effort was a project called Rex, a real estate search and metrics tool (hat tip to colleague, Steve, who originally suggested the 10% time idea to me back in 2003) . This project provided a way for people to combine data about the real estate market with a mechanism for prioritizing the types of places they’d like to live. It required us to develop two innovations: 1) a way to project likely home prices based on recent market sales; and 2) a way to perform very fast raster data processing in order to prioritize the best locations.
Our attempt to turn this into Azavea’s first product was a failure, but we learned a lot, and our technology innovation was successful. We turned this into an economic development siting tool (Decision Maps) that we built for the City of Philadelphia. In turn, that became a grant proposal to the US Department of Agriculture, which was our first ever Small Business Innovation Research (SBIR) grant proposal. We were fortunate to receive funding for it, and it gradually became a more generic toolkit for processing raster data. This later became the conceptual starting point for GeoTrellis, Raster Foundry, and many other raster processing tooling that we can continue to use and build upon. Geospatial imagery and other raster data processing is now a key area of business for us.
Redistricting and Gerrymandering
The United States has a significant gerrymandering problem at every level of representative government. Azavea has been actively developing tools to support more open, citizen-driven redistricting for the past two decennial census cycles. However, our work in redistricting did not start with software development; it began with a 10% time data project by Megan, a GIS Analyst. Megan had a question: Philadelphia’s City Council districts looked incredibly convoluted and were clearly an example of gerrymandering; but, if we compared Philadelphia’s districts to other cities, just how bad were they? Megan used her 10% time to research and compute compactness metrics by leveraging our Cicero database. Then she turned the results into a white paper on gerrymandering and compactness metrics. In 2009, just before the 2010 Census, we began working on a revised series of white papers as well as thinking about what a web-based redistricting tool would look like. This grew into a long term partnership with the Public Mapping Project, led by Dr. Micah Altman and Dr. Michael McDonald. Last year we relaunched version 2 (the 2020 edition) of our free and open source redistricting tool, DistrictBuilder, for the current redistricting cycle.
Use of Amazon Web Service (AWS)
In 2022 it seems hard to imagine a time when software companies were operating entirely on physical servers. Azavea’s shift to cloud computing infrastructure began as a 10% time project by colleagues we affectionately referred to as The Two Daves. In late 2006 (relatively early days for AWS), they were both interested in experimenting with what this emerging technology could make possible. They each experimented with AWS and shared back with the company what they learned. In most cases we don’t necessarily make big changes from a single 10% time project, but in this case we soon began shifting our work to public cloud infrastructure and are almost entirely reliant on it today.
Contributing to open source
Rich was the first Azavean to explore the use of open source geospatial tools. When Azavea started we mostly relied on ESRI and Microsoft .Net technology. Rich channeled his curiosity about emerging open source geospatial tools into first testing open source tools and then building a prototype project, the Culture Browser, a web app that combined data from PhillyHistory, the Mural Arts Program, historic markers, and other open datasets, to build a mobile phone app for exploring the cultural sites in our city.
This was the kernel of open source at Azavea, and it opened everyone’s eyes to what we could do with open source technology. Over time, our clients began requesting open source tools and frameworks. Esri and Microsoft technology remain important parts of the technology ecosystem, and we rely on them in many scenarios, but open source tools are now the primary building blocks with which we deliver applications to our clients.
Other 10% projects that represent contributions to open source technologies include:
Pro bono projects
Some of our colleagues apply their 10% time to developing new tools for nonprofit organizations. One recent example was TransitAnalyst, a collaboration between Kathryn and Mike that used open federal data sets to create an online application that combined public transit data with housing and economic data to help communities understand where critical resources are located relative to transit access.
Learning projects
The 10% time program is not just for building new tools. The core objective is to encourage everyone in the organization to set aside time to learn and develop professionally. Building or improving technology is a way to do that, but we also encourage people to read papers, attend conferences, and participate in workshops. Reading groups are another popular use of 10% time that have formed around topics such as climate change, functional programming, machine learning, marketing, product development, and people management.
The value of 10% time
I’ve highlighted several successful projects above, but it’s important to note that most 10% time projects don’t change the course of the company, and a non-trivial number are abandoned mid-stream. Between those two extremes – changing the organization’s direction and being abandoned – some are turned into utilities that we use across the company (like the GTFS tracker puller), while some are exploratory (like Google Glass). I think part of what has enabled the program to be successful over time is the basic assumption that many efforts won’t be directly fruitful for the organization, and that’s to be expected. However, even when an effort “fails”, in the process, we are exploring the space of what’s possible, and that is always valuable.
Innovation can occur in any environment, but I think significant, ongoing innovation requires a combination of structure and an element of chaos. Too much structure – “this is the way we do things and we’re not changing it” – or a top-down identification of “valid” new ideas by executives will not be conducive to the organization adapting to change. Conversely, a lack of structure in which anything goes, won’t allow significant changes to be developed and propagated across an organization. The happy medium is a balance between structure and chaos: 1) an open invitation and time allocated to try our new ideas, and 2) a willingness to both double-down on new ideas that have traction and walk away from ideas that don’t show promise. Innovation is part of Azavea’s mission. Our 10% time program is not the only way that we pursue that mission, but it is one that helps us try to strike that balance.