Wisconsin’s elections are fully geo-enabled, after a conversion process that started in 2011. The state had concluded that implementing GIS would make the process of assigning voters to the right districts, following the 2010 census and redistricting, more efficient and accurate.
The Wisconsin team’s biggest challenge by far turned out to be the quality of available data – both for voter addresses and for district shapefiles (the geospatial data format that defines geography). While data has improved significantly over time, finding sources of high-quality data to use for auditing and quality control remains a concern, and the team shares some of the approaches considered and used.
The initial step towards integrating GIS in the election process involved running all existing addresses through a commercial geocoder to determine their X and Y coordinates. Next, by comparing this spatial data to district maps, certain issues were automatically flagged for review. While this created a considerable workload for local officials, there were also potential issues that the system at that time was unable to automatically identify, such as voter address points located in the middle of a street.
Today, the state’s systems have improved. Also, the team is working towards a program of frequent audits – as often as every six months or yearly – where existing entries are automatically compared against high-quality external data to highlight any possible issues.
Based on their experience, the Wisconsin team recommends asking the following questions before starting the process to implement GIS in elections:
- Do you have the data? If not, where are you going to get it? Is it high quality, or how will you improve data that has shortcomings?
- What is your plan for regularly auditing and improving data over time, using a reliable data source?
- What are the possible unique situations that will cause problems for your system? For instance, some colleges in Wisconsin have a centralized mail facility for students, while the students’ actual living accommodations may fall on different sides of a district boundary.
A final challenge identified by the Wisconsin team involves increasing the understanding among all stakeholders and data providers of the importance of accuracy in assigning locations to voters and finding ways to get all stakeholders to buy into the idea of improving to the process.
On a positive note, there may be opportunities to collaborate on a state level. Other agencies may have databases that election officials can tap into, and, conversely, the improved data of election offices may be helpful to other agencies. Also, there may be statewide or local efforts in place to bring together databases that can aid in geo-enabling elections.
In conclusion, GIS uses spatial logic to verify that voters are in the correct place, instead of relying on spreadsheets or only local knowledge, improving overall accuracy and efficiency in elections management.
Read Wisconsin’s case study in more detail here.
If you missed the first case study from Utah, click here to read additional findings.