Mark Elbert has served as the director of the Office of Web Management for the U.S. Energy Information Administration for the last four years, maintaining and enhancing the website. For information contact him at mark.elbert@eia.gov.

The U.S. Energy Information Administration (EIA) is using the geographical attributes of its statistical data sets to create new web products with interactive maps, highlighting the geospatial component of the nation’s energy infrastructure and resources. Two recent EIA web products, the State Energy Portal and the Electricity Data Browser, are exemplars of the potential to add data-driven maps to existing publications.

The State Energy Portal contains information such as rankings, key statistics, quick facts, analysis, and links to additional resources for each state. However, the interactive states maps, made possible by EIA’s new energy mapping system, steal the show. Users can choose among five base maps, such as satellite imagery or a street map, and then layer on any combination of 30 available layers, such as natural gas fields, electric power plants, federal lands, and renewable resources. The application also has 12 predefined themes such as infrastructure, fossil resources, or electricity to simplify the layer-selection process.

The second product, EIA’s Electricity Data Browser, has a number of visualization tools, including animated heat maps of the United States, to visualize state-level statistics over time. While the Electricity Data Browser’s maps are simple infographics compared to EIA’s national mapping system, they do display a temporal component. EIA launched both products in the past year during a period of rapid innovation in the area of web-mapping tools.

Emerging capabilities of infographic map tools, such as the one used in the Electricity Data Browser, will blur the line between interactive infographic maps and true geographic information systems. The key to leveraging any of these mapping tools is to have well-structured databases with defined geospatial and temporal attributes.
Both products used newly released tools for creating interactive web maps and graphs, but took different implementation approaches, each with their strengths and weaknesses.
The Electricity Data Browser was developed first. It was custom written, but employed 20 open source JavaScript libraries to perform specific functions. For example, the Highcharts library creates the interactive charts, while the jVectorMap library creates the heat maps, which are animated with the help of the jQuerylibrary. This approach allowed EIA to produce and launch the first version in just 12 weeks using a team of two programmers, a graphic designer, and a manager. Examining the browser’s code base, the open source libraries comprised about two-thirds of the total code and were instrumental in the project’s quick delivery. The team had the freedom to pick which open source libraries to include, creating a unique platform of tested and documented building blocks. This limited the custom programming to interfacing between the open source components and the data services on EIA’s server to respond to user queries.

EIA’s state portal maps were developed immediately afterward. The portal uses ArcGIS Server, a commercial off-the-shelf (COTS) mapping software by ESRI, to implement its online maps. ArcGIS Server delivered configurable maps supporting five base maps and any combination of 34 energy resource, infrastructure, and administrative boundary layers. Setting up the server was relatively simple. The bulk of the work was actually gathering, creating, and checking the layers.

With such different approaches, it is natural to ask which is better.

Before discussing the pros and cons, it is worth noting that online mapping is a space in flux. The creators of the various mapping servers and toolkits are rapidly innovating to take advantage of the computing and rendering capabilities of the latest web browsers. This means the gap in mapping capabilities between the commercial leaders and the open source community is narrowing. New open source web-mapping libraries such as Leaflet.js and MapBox are emerging with impressive new features. Therefore, selecting an enterprise mapping solution for the web should be avoided at this time. It would simply be unwise to cut off your organization from this innovation. Business needs will likely dictate different map solutions for different web products, as the mapping solutions tend to be optimized for different mapping tasks.

When maps are building components of a statistical information system, the open source approach has clear advantages. In statistical information systems, the need is often for less detailed maps to visualize aggregated statistics as heat maps (also known as choropleth maps) or to provide sized and color-coded markers to visualize statistics associated with point locations such as the electricity generation of power plants. Open-source libraries can be combined easily with charts, tables, and user controls to create powerful statistical query tools.

The Electricity Data Browser, for example, allows users to find, chart, and analyze any of 430,000 time series. State-level aggregate statistics can be visualized as a heat map of the United States, and even animated to visualize patterns over time. In every instance, the data and relative standard error can be inspected and downloaded and aggregated statistics can be drilled down to the underlying generation plants’ data. Effectively, using open-source building blocks makes the development of advanced custom query tools like the Electricity Data Browser affordable.

That said, we continue to be pleased with our selection of ArcGIS Server to power EIA’s complex state portal maps. We found no open-source solution that could handle more than 30 interactive layers. To reiterate, different requirements drove the selection of different mapping tools.

COTS solutions offer an easier route to mapping statistical data. The maps are interactive and have impressive layering capabilities. However, COTS software lacks the flexibility to combine maps with custom query tools in truly novel ways to present a particular data set optimally.

Open source is often better for a full custom solution and for data-driven infographics. EIA has found open-source libraries to be capable and only getting better. The implementation time will be longer and require a highly skilled web development team. However, a small team with the right programming and presentation skills can create cognitively advanced web products designed to best present a particular data set.

Therefore, for the foreseeable future, we expect to use a variety of tools and approaches to visualize the geographic attributes of the agency’s statistics, depending on a project’s goals and the evolution of the available visualization libraries and toolkits.