By: Eric Gundersen
Crossposted from VentureBeat.
When truth is increasingly subjective, maps can be a canvas for people to understand data, and through that understanding drive faith and confidence in how government is elected.
A U.S. election is an incredibly complex data pipeline. Who can vote is determined by fifty-six different sets of eligibility rules and eligibility lists. To vote, individuals must physically access the polls or request a ballot. They receive countless versions of ballots and record their preferences using pen, hole punch, touch screens, and more. The ballots are counted by different types of machines and sometimes by people. Counting rooms are located in thousands of different locations. Results are aggregated from small local districts up to full states. It all boils down to two or more numbers per race — who won, who lost. These numbers determine our leaders, our laws, our culture, our future.
Integrity, equality, accuracy, and security are vital — not only to running an election, but to the fundamental tenet of democracy, the trust in the system that our voices are heard and counted. How do we evaluate a system that is so complex? Location and data provide essential context to understand the results of an election and the functioning of our democracy.
I saw this first hand in Afghanistan in 2009, mapping the election results with State Department and National Democratic Institute. When we got to Kabul, the maps were blank. How were we going to map the election? We had to build the map, and in doing so started a company.
At that moment we were trying to understand what had just gone wrong in an election, we weren’t trying to create a mapping platform. Election day — August 20, 2009 — was the most violent day in Afghanistan since the Taliban was ousted. Participation and logistics were complicated. Yet many precincts were reporting 600 ballots — meaning 100% of all possible ballots were cast — and/or reporting 95% or higher support for a single candidate.
The patterns of the irregularities and outliers were hidden in two enormous PDF documents — the list of 6,969 polling centers and the Independent Election Commission’s (IEC) 2,500-page PDF of preliminary election results. The ability to put the results on a map, and compare with additional layers of information — including the location of security threats and the primary ethnic group by district — revealed the extent of the fraud, where it occurred, and that it clearly benefited one candidate. The IEC audited 3,300 polling stations and over 1.5 million ballots. Given the findings, 19% of all votes cast on August 20 were excluded from the final presidential vote tally. This report, our maps, and data that we released in open formats, is still available at AfghanistanElections.org.
This has scaled tremendously. Today it’s inspiring to see different organizations using our mapping platform to bring new understanding of our elections: Projects like the ACLU’s “What The District” — which visualizes changing congressional district boundaries, affecting our representation and even how our votes are counted — The New York Times national precinct-level map of the 2016 Presidential Election — which reveals a mosaic of varying political support across the country — and ProPublica’s report mapping how early voting changes in North Carolina are impacting access to the polls for voters in the state.
In the long term, we must modernize voting systems, insure equal access, and secure election infrastructure. In this moment, using mapping tools to provide context and analyze the data is more important than ever.
In Afghanistan, that election map changed the conversation. A map answered the question of whether we really needed a runoff. Today, we can look at the complex election data pipeline with new clarity — to understand where it’s functioning as it should, and where it isn’t.
There’s a vital role to play for the democratization of mapping tools, and of the data sets that power them. We know maps can be used to inform, and to mislead, but if the data has integrity then so will the map. Let’s start doing more data-driven debating and exploration. Our democracy depends on it.
If you’re working on mapping election data publicly, Mapbox will offer free use of our platform and tools for media and advocacy organizations. Check out our team’s guide to election mapping to get started. And drop me a line at eric.gundersen@mapbox.com and I will loop in the right people to make sure your account is activated.
Adding truth to the election debate was originally published in Points of interest on Medium, where people are continuing the conversation by highlighting and responding to this story.