In last week’s post, we detailed some of the ways political campaigns use GIS and spatial information to bolster their efforts. Just as contenders in those contests benefit from diverse data, so too can parties responsible for organizing and running those elections. The goal? To make voting freer, fairer, and more accessible.
In much the same way geo-tagging voters helps campaigns identify areas and addresses to target, the people who organize can use spatial data to communicate information, place voters in correct districts, and set up enough polling places to satisfy voter demand (thereby avoiding long lines at the polls). In 2019, NSGIC released a guide for GIS-enabled elections, which they made available here. The document outlines five best practices, which consist of hiring a team of specialists, creating and maintaining a voting unit GIS layer, developing a statewide geocoding strategy, assembling a contextual GIS layer, and implementing a data validation process.
Those practices are aimed at increasing voter turnout, making participation easier, and clarifying who should be voting where. At times, however, those goals seem at odds with existing districting systems.
In the US (though the issue persists in other representative democracies), “gerrymandered” districts are something most voters can agree are unfair. Gerrymandering is the practice of manipulating district boundaries to tip the balance in favor of one party or another. In the US, efforts to stymie gerrymandering have been met with resistance from law makers, despite the practice’s unpopularity. The result is an electoral map filled with shapes only an incumbent could love.
Diagram courtesy of M. Boli, CC BY 4.0
Luckily, spatial data has no political affiliations or aspirations, and leveraging its objectivity is a way to increase the fairness and accessibility of elections in systems like that in the US. And with 2020 census data on the horizon, a new way forward is possible.
As detailed in this article from GIS Lounge, there are specific indicators that point to an un-tenably gerrymandered district. Voting history and district’s boundaries are compared against data like demographic breakdowns, landscape features, income averages, metrics of shape, and other local factors.
A gerrymandered district in Illinois. Image courtesy of CBS News.
This study by Kamyoung Kim of Kyungpook National University’s Department of Geography Education details a specific approach for optimizing spatial data to detect biased districts, called the capacitated double p-median problem with preference (CDPMP-P) approach. This approach analyzes a district’s spatial data, population relative to representation, and whether or not the voting infrastructure is robust. If the area scores high, it is likely not gerrymandered. A low score, however, means that district may need a democratic overhaul.
With these tools, election officials and organizers can help shepherd electorates around the world into a fairer democratic future.
In previous posts, we’ve covered how governments utilize GIS and diverse spatial data to respond to crises and plan for future ones. In nations with electoral systems like the United States, people who want ‘in’ to the government utilize much of that same information to run their campaigns and spur voter turnout.
A visual representation of every vote cast in the 2016 United States presidential election. Image courtesy of Kenneth Field and Esri.
Since the early 2000s, (or at least as far back as 2004) political campaigns have added ‘microtargeting’ to their tool kits – essentially turning direct marketing methods into outreach tools for campaigns.
In this post from 2008, Karsten Vennemann of Terra GIS describes his experience working for the Obama campaign. Back then, he made use of open-source technology to identify and target specific voters. Put simply, the process allowed the campaign to lay voter data over base layers like precinct and county boundaries. The approach was so well liked, according to Vennemann, it was implemented in roughly a dozen highly contested states.
Today, tools like those offered by Esri’s ArcGIS have streamlined that process of turning addresses into useable data points. The first step is turning a list of addresses into a ‘geocoded’ list of voters. To do this, tools like Esri’s Tapestry Segmentation will analyze neighborhoods by socioeconomic and demographic factors. The program then classifies neighborhoods using 67 categories of market segments based on information like income, hobbies, education, and religious affiliation. Combine the Tapestry information with a list of addresses, and a picture of potential voters and donors emerges.
Esri Tapestry information of potential donors & voters. Image courtesy of Esri.
When that data is combined with information about boundaries and districts, campaigns can begin formulating ground plans. They can focus their attention on areas with potential support, while also avoiding areas where their cause may be hopeless. Essentially, this eliminates figuring out where to door-knock the hard way.
In today’s world of online data, evolving voter habits, and other factors, the ways in which this approach can be applied will likely continue to evolve. In next week’s post, we’ll take a look at how governments can use these same tools to ensure fair, accurate elections.
Doing business in the era of COVID-19 means navigating around office closures, travel restrictions, and other roadblocks. For cartographers & GIS teams, who often work in a collaborative environment, and by leveraging large datasets sitting on local servers, this reality could present major problems.
Thankfully, current technology and strong, pre-existing relationships between data providers and map publishers worldwide allow for remote work where it was once not feasible. This has kept an essential industry running in uncertain times.
The East View Geospatial cartographic team found itself at the forefront of this phenomenon earlier this year when it produced mapping of the Polynesian island nation of Samoa.
Just prior to its work in Samoa, the EVG production team published 16 topographic maps over Tanzania after identifying gaps in the series’ coverage. This effort thereby helped complete the Tanzania 1:50,000 topographic map series for the first time in the nation’s history (this story map tells that story).
In their efforts to execute the map production efficiently, accurately and up to MTM Mil-Spec requirements, the team utilized a suite of workflow enhancement tools developed by the Esri Defense Mapping Team, just as they had in their Tanzania work. These tools streamlined the production process and the team’s expertise with said technology made the completion of the Samoan map series possible.
One of the topographic map sheets over Samoa produced by the EVG cartographic team.
In the end, our cartographic team produced & published a series of 1:50,000 scale topographic mapping comprised of 16 map sheets over Samoa. This series is the most recent addition to the EVMap catalog that includes datasets over Tanzania, Nigeria, Mexico and Myanmar. The production of this map series was executed 100% remotely and completed in 50% of the time estimated.
Mapping in the era of COVID may come with complications, but with the right tools in the right hands, precise, ground-breaking mapping projects can seamlessly move forward.
Watch the short video below to see how the EVG cartographic team mapped Samoa in the age of a pandemic:
There is more data in the palms of our hands than ever before, and even for industries built on selling access to now-free information, there is a silver lining. Access to publicly available data is heightening everyone’s imaginations, and, in doing so, is stimulating demand for further expertise in data science and technology to harness its full potential.
For specific outcomes, though, expertise is needed. Many companies and individuals, while cognizant of this open data revolution in which we find ourselves, still need assistance in both sourcing the appropriate data as well as leveraging it to reap intended outcomes. To illustrate the demand for and value of data science and technology expertise, detailed below are recent projects leveraging open-source data.
Building height data produced by EVG, shown in red, compared to real building size to show accuracy.
Satellite and aerial imagery represent another data type that has become exponentially more pervasive in the public domain, and these remotely sensed datasets can enable the creation of innumerable derivative products and applications. Good examples of these include satellite-borne images from the Landsat Missions, made available through NASA and USGS, and Sentinel-2, made available through the European Space Agency. Many are unaware of the existence of such imagery. Or, even if they are aware, they may struggle to exploit the data into meaningful outcomes. That’s where solution providers, like East View Geospatial, come into play.
In another notable case, EVG was able to utilize Landsat imagery over Costa Rica from 2015-2018 to perform a detailed change detection study on behalf of our client. This publicly available imagery allowed us to quickly analyze land cover and canopy changes over the country during a time when Costa Rica had experienced two major hurricanes. The data outputs we provided allowed the client to evaluate the extent of the environmental impact as well as their post-disaster response.
Satellite imagery over Costa Rica compared to land use/land cover data over the same area.
In both cases, EVG’s knowledge of public data & access to difficult-to-source data through National Mapping Agencies combined with our technical expertise enabled an agile solution. Aggregating free data, analyzing, conflating it with other datasets and delivering on a plan of action is something that has become integral to our operation since its inception. Growing access to free data has ushered out many traditional models of selling data but, most importantly, it has presented new opportunities to bring forth creative solutions to today’s most difficult challenges.
As complex spatial data and GIS have grown more and more ubiquitous, so too has the public’s ability to better comprehend events happening time zones away from them. Likewise, the rise of such technology has also improved people’s ability to react when disasters strike close to home.
In August, a deadly explosion in Beirut captured the attention of the globe. Footage of the event spread rapidly over social media, but for many, understanding the magnitude of a blast in a city they have never visited was made possible with diverse visualizations like those in a New York Times article. In one image, circles representing the blast radius are overlaid on top population density data. In another, satellite imagery of the city is overlaid with data points like kinds of damage reported (from leveled buildings close to the explosion to blown-out windows miles away from its center) and other information specific to individual neighborhoods.
Image courtesy of the New York Post.
Other large-scale natural disasters, like wildfires, hurricanes, floods, landslides, and air pollution, are often difficult for individuals to truly comprehend but also necessary to react to, in both the immediate and long-term. GIS allows multiple forms of data to overlap and paint a clear picture of both the physical and socio-economic impacts of a disaster.
In the United States, the Federal Emergency Management Agency’s Hurricane Incident Journal keeps tabs on both the severity of an approaching storm and how much potential harm already-disadvantaged populations may face. National Oceanic and Atmospheric Administration (NOAA) data projecting the path of storms is conflated with data from the CDC’s Social Vulnerability Index. For 2020’s Hurricane Laura, COVID-19 case data from Johns Hopkins University was also added to that mix. The United States Army Corp of Engineers’ Flood Inundation Mapping project provides an interactive map with levels of information from stream gauges and levee information to help predict where floods may strike afterward.
Image courtesy of WDSU News.
Similar tools exist for tracking wildfires, like those currently ravaging the west coast of the US. The United States Geological Survey’s (USGS) Landfire Data Distribution Site tracks fires in real-time, as do several other non-state organizations. Advances in unmanned air vehicles and the Internet of Things have allowed for more rapid updates with less risk to humans.
Recently, Google has added data to ensure that these kinds of apps can continue to be frequently updated and rapidly accessed, meaning the tools stay useable even as demand for them increases with the advent of a disaster.
After fires burn through the vegetation that, quite literally, holds land together, landslides are a common occurrence. The United States Geological Survey (USGS) has tools to monitor those threats in real-time, including landslide monitoring and post-fire debris flows, and informational guides for understanding the maps. The USGS also provides similar tools for earthquakes. This data is created in cooperation with institutions like the Colorado School of Mines.
Image courtesy of The San Francisco Chronicle.
Of course, since this is spatial data, there are non-governmental actors providing knowledge as well. One example is PurpleAir, an air quality tracking site created using the Internet of Things, including devices you can buy on-site to monitor air around your home and contribute to the overall data. For a non-governmental live look at developing storms, Esri’s disaster response hub houses a wealth of tools.
Geospatial data continues to play a crucial role in not only tracking & responding to a natural disaster but also informing the public on the gravity of the situation. The tools and visualizations made available by geospatial data and GIS software continue to provide key insights and information for decision-makers and laypeople alike.