Building a Web GIS Business


I’d like you to meet an interesting Penn
Stater. Stephen Ansari is co-founder and principal of Blue Raster, the company that you heard about in the video prelude to this case study. This story is about how Stephen and his partner built a successful business in Web GIS. It’s also about what you can do to create opportunities for fulfillment in your
own career. I met Stephen in 1997, when he was an undergraduate student in a course I taught at Penn State. In fact, that course, called Mapping Our Changing World, was an early on-campus version of the online course you’re taking now. In those early days of the web, I built HTML files to illustrate and share my lectures, and assigned students to code HTML too. Stephen came to Penn State from Downingtown High in southeastern Pennsylvania to study earth science and mapping. He was already into web programming by the time he enrolled in my class. Stephen had been coding since age 8, when his family bought an apple IIe, and he found a book on Basic programming. Later he upgraded to a Tandy 1000 SX, which he says opened up a whole world of programming simple applications and user interfaces. He says he loved the immediate gratification of building things with code. It was like engineering but without the need for raw materials. For a time coding remained a hobby for Stephen. His academic interests centered on earth science and mapping. Then in 1995, the US Census Bureau launched its TIGER Map Service, one of
the earliest web mapping services. Even with his slow modem connection to the Internet, Stephen was amazed at the ability to explore the Census Bureau’s
TIGER/Line files and thematic maps online. Later in our Mapping Our Changing
World class, Stephen discovered the United States Geological Survey’s Digital Raster Graphics. As you know, DRGs are topographic maps that have been scanned
and georeferenced. That’s when he started to see connections between maps
and computers. As an earth scientist, working on the field, he’d always thought
of topo maps as just printed things. Maps and the world really were changing. But it wasn’t until his first job out of college that Stephen got to put these insights to work. After graduation in 1999, Stephen found a job in the Washington DC area as a policy analyst for a firm that was writing
groundwater guidance documents for the US Environmental Protection Agency.
Working at the desk next to him, was Michael Lippmann. Mike had earned a
bachelor’s degree and then a Master of Public Administration and Environmental
Policy at Cornell. A fellowship at EPA brought him to DC where he later found
work as a Policy Analyst at the same firm as Stephen. The co-workers shared an
interest in the potential of web mapping and an early Internet Map Server
technology called ARC IMS, which their employer was just beginning to
experiment with. The company sent Mike and Stephen to Redlands for training in
the new technology. They made connections there that would serve them well later. Some of their employer’s government agency clients wanted to begin mapping contaminated sites in the US. That provided Mike and Stephen’s first foray into GIS. Although their employer didn’t create the National Institutes of Health toxmap shown here, the example does suggest what an ARCIMS-like web mapping app looked like in those days. Stephen and Mike began to envision how web GIS could be a game-changing technology. But electronic maps still seemed esoteric to most of their employers clients. Even Google Maps was still years in the
future and the dot-com boom had recently gone bust. It would take a while longer before a business model based on web mapping and GIS would seem viable. In 2002 Stephen and Mike took the plunge and founded Blue Raster, LLC. In interviews for this case study, Steven and Mike explained that the name “Blue Raster” is meant to connote the ability to capture complex reality systematically and to represent it simply, in ways that reveal the meaning their customers wish to convey. That’s the value they aim to deliver with Web GIS. Blue Raster’s earliest contracts involved creating web mapping apps about radioactive contamination at research sites for the Department of Energy. The first of those sites was Brookhaven National Laboratory on Long Island. Brookhaven operated a “High Flux Beam Reactor” used for scientific research from 1965 to 1996. According to the Department of Energy, “The reactor was decommissioned in 1996 after tritium, a radioactive form of hydrogen and a byproduct of reactor operations, was found in nearby groundwater monitoring wells.” Investigators subsequently found that contaminated water, quote, “had been leaking from a spent fuel pool located in the basement of the reactor building for more than a decade.” As urban development spread eastward on Long Island, public concern about contaminated groundwater increased. Years before the term “open data” was coined, Brookhaven commissioned Blue Raster to create web apps that provided maps and other site-specific information graphics
needed to keep residents and the press informed. For example, users could drill down into monitoring well locations to view graphs like this, that showed how tritium concentrations decreased at the wells with time and with restoration efforts. Functionality like that may seem obvious now, but it wasn’t then. Which leads me to wonder, what will users expect from web mapping apps 10 or 20 years in the future? Brookhaven engineers and contractors
modeled and tracked the Tritium plume and shared the data with Blue Raster for
the website. Every week Stephen and Mike would take an early morning flight from
Baltimore-Washington International Airport to Islip, Long Island. After
consulting with the engineers they’d travel home to Washington on the last flight of the day. Brookhaven National Lab met its
remediation targets and Blue Rasters public outreach was considered a success.
Similar Department of Energy projects followed. In time they were branded
programmatically as LandTrek, but Brookhaven’s success in meeting its goals
convinced Mike and Stephen that Blue Raster needed to diversify. They experimented with maps for election campaigns and apps for the real estate
market. They got their name out by participating in meetups in the DC area. Finally one of their contacts at Esri introduced them to the World Wildlife Fund. The World Wildlife Fund had compiled a database of the global distributions of thousands of mammals, amphibians, reptiles, and birds. In 2004, WWF hired Blue Raster to design and build a web mapping app called WildFinder, that lets users search the WWF database and visualize ecoregions in
which species live. Mike explains that they built the first version of the web app with ARCWeb Services, the technology that evolved later into ArcGIS Online. In 2009, Blue Raster and WWF released WildFinder II, which is built on ArcGIS for Server and Flex. To keep pace with evolving technology Blue Raster hopes to create an updated version built with the ArcGIS API for JavaScript. Working at the intersection of conservation and global affairs, Blue Raster began to hit its stride. The WildFinder app became, for a time, one of the most visited areas of the World Wildlife Fund’s site. Their success caught Esri’s attention. and just four years after its founding, Blue Raster was named Business Partner of the Year, in 2006. Remember this conceptual map of the “Web GIS technology ecosystem” from an earlier case study? Mike Lippmann co-authored it. Like his
collaborator Robert Cheatham, Mike stresses that this should be considered
a living document that’s updated continuously. In the lower middle of the
map, notice the cluster of technologies labeled “map servers”. One of Blue Raster’s
core competencies is the ability to deploy server technology, as a means to
publish geospatial data and interactive geo apps on the web. One example of a project powered by server technology, is Global Forest Watch, an initiative sponsored by the World Wildlife Institute. Global Forest Watch includes a suite of data, maps, and apps all intended to help people everywhere to better protect forests. One of those apps is Global Forest Watch Fire. Show here is Blue Raster’s 2014 wireframe design of the Global Forest Watch Fire analysis interface. Driven by concern about the role of fire and deforestation and air pollution, WRI convened a partnership involving Blue Raster, Digital Globe, NOAA, and others. to create a high-resolution, near real-time, global fire monitoring system that anyone can access and use on the web. This is a design comp, a next stage in the app development process, at which the customers vision for the
product and the developers’ vision begin to converge. The live version I viewed in July 2017 reflects expanded functionality, including crowdsourcing. Global Forest Watch Fires brings together several organizations and technologies. The MODIS satellite-based sensor is used to detect fires around the world. Daily updates of fire events are stored and served from ArcGIS for Server. High-confidence fire events trigger the capture of high-resolution DigitalGlobe imagery, which is made available for near real-time viewing and analysis by the ArcGIS image Extension for Server. Global Forest Watch users can subscribe to fire alerts which are delivered by email or text. And to help visualize the potential impact of air pollution, an open-source JavaScript routine is used to create animations of wind directions and speed every four hours. Blue Raster uses either ArcGIS
Enterprise, which Esri called ArcGIS for Server until version 10.5, or ArcGIS Online, to serve their web maps and apps. The choice depends on project
requirements. Other options are available, of course, including Map Server, GeoServer,
and others. You’ll have a chance to explore the alternatives if you take the
Penn State course, “Cloud and Server GIS”. The ability to “design a geospatial
system architecture that responds to user needs” is a key competency for Blue Raster, and maybe for you too. In the video prelude to this case study,
Christine Movius pointed out the importance of coding in a web mapping
business like Blue Raster. Of the several languages shown in the Web GIS
Technology Ecosystem map, Python and its associated libraries, are perhaps most
important for GIS pros. Python is useful for automating repetitive tasks, for customizing desktop GIS interfaces, and for what data scientists call “data
wrangling”. A common estimate is that up to 80% of data analysis is consumed with
preliminary tasks of cleaning and preparing data. Python is widely used for
such tasks. But as Python has matured, geoprocessing has emerged as another common usage. Libraries like PySAL (the Python Spatial Analysis Library), and Application Programming Interfaces like the ArcGIS for Python API, make Python an
increasingly valuable tool for serving what Mike Lippmann calls “an increasing
demand for more robust spatial analysis.” One Blue Raster project that benefited from Python scripting is LandMark. In collaboration with the World Resources Institute and others, Blue Raster designed and built LandMark as the first global repository of community land rights data. According to WRI, as much as 65 percent of the world’s land is held by indigenous peoples and local communities, but little more than 10 percent is recognized as belonging to them. The LandMark platform allows users to draw a shape on a map to query the attributes of communities within that area of interest. Blue Raster developer David Eitelberg used Python to “calculate the quantity of indigenous and community lands within political borders.” LandMark is useful precisely because it combines many disparate data sources, in forms that allow further analysis and visualization. Stephen Ansari says that ten years ago GIS pros used desktop software to perform analysis and then produce derivatives that might be included in web Maps. Today, he says, we’re using Python to build repeatable systems that run in real time and can do on-the-fly combinations of data that are controlled by end-users. He urges every GIS pro to hone their Python skills. Python scripting is specifically called out in the GTCM. Opportunities for python training abound, including Penn State’s online course “GIS Programming and Customization.” The 2015 Paris Climate Conference was the 21st annual “Conference of Parties” at which representatives of 195 nations meet to assess progress in confronting and reversing human-induced climate change. In the lead-up to the December conference, hopes were high that an international agreement to keep global warming below two degrees Celsius was possible. A mere two months before the Conference, Blue Raster took a call from UNICEF, the United Nations Children’s Fund. Believe it or not, UNICEF discovered Blue Raster, out of the blue, with a Google search. UNICEF wanted to produce a compelling statement about the risks that climate change poses for the world’s 2.3 billion people under age 18. They had lots of data and analysis but no clear idea of how to tell a story that would make a difference in Paris and beyond. Stephen, Mike and Kevin McMaster engaged their clients in an iterative design and development process that soon yielded
information products that helped crystallize UNICEF’s story. For example, this map of the distribution of children living in flood occurrences
zones combines UNICEF analysis with flood maps produced for the Aqueduct online
water resource atlas, which Blue Raster developed with the World Resources
Institute. Blue Raster succeeded in helping UNICEF complete its publication, “Unless We Act Now”, in time for the Paris conference. Later Stephen and Mike called it their “proudest achievement” to date. The experience in contacts they gained
brought more collaborations with UNICEF, including “Clear the Air for Children” in
2016 , and “Thirsting for a Future” in 2017. Following the “Thirsting for a Future”
project, UNICEF policy analyst Nicholas Rees said,
“Blue Raster has helped us make the case for protecting children’s access to safe
water and sanitation.” A common thread that connects Blue Raster’s projects is
their emphasis on helping clients define goals that are consistent with their
organization’s purpose and in crafting stories that advance those goals. To support Blue Raster’s growth, Stephen and Mike look for employees who are “good communicators” – who can “geo-enable their customer’s” data to “tell stories with abundant information.” Here they stand with Penn State alum Chris Gabris and Christina Phang, who co-hosted my first visit to their offices in 2016. Chris recently developed the U.S. National Arboretum’s Botanical Explorer mobile app, which helps visitors and researchers explore the Arboretum grounds and plant collections. Among many other things, Christina created the “Story Maps Starter Kit” we use in this course. I asked Stephen and Mike how they determine whether an applicant has what it takes to be successful at Blue Raster. Their job interviews “start with non-technical conversations” about “what excites” applicants. They look for “innate talent, good imagination, writing skills, public speaking ability”, and especially excitement about geography. Next they look for tech skills such as programming, the JavaScript API and newer JavaScript libraries, HTML5 and IT skills such as server setup and administration. Mike remarked that they look for people who fill the gap between college curricula and the real world. They also seek applicants who are “active in the community” as in dev meet-ups and user groups, and who assert “thought leadership on the web.” Finally and especially, they need people who “know how to find things” – who have the right mindset to learn collaboratively and independently – and who are “energized to move on to the next level.” These same traits that prepared Stephen and Mike to build a successful business in web GIS are what we hope you will develop and build upon in your time with us at Penn State.

, , , , , , , , ,

Post navigation

Leave a Reply

Your email address will not be published. Required fields are marked *