Morgantown 2021 Municipal Election GIS Report

The Morgantown Municipal Election took place on Tuesday, April 27th, 2021. This was the first Municipal Election cycle that saw involvement with the City's GIS implementation, and therefore presented a great opportunity for improvements of workflows but also visibility of geospatial technology's use in local government. With the completion of the Canvass on Monday, May 3rd, 2021, the election cycle has come to a close and this report will detail the solutions deployed, the solutions utilized, and an evaluation of those solutions.

I approached the City Clerk, Christine Wade, with a few ideas during the Ward Boundary Analysis that took place in mid-to-late-2020. Primarily, the ability to report results to a Dashboard was what was discussed, but once I explored all the Election-related solutions available through the ArcGIS Solutions platform, there were several that the City Clerk wished to implement for the Election. The solutions that were deployed and utilized are shown below.

The Voter Satisfaction Survey had a response rate of 181 out of 1,534 ballots cast (11.8%). The Dashboard below provides a quantitative breakdown of the survey responses and provides the user the ability to filter the results by voting location.

The general response for the implemented solutions that was received from the City Clerk and those who worked the polls during early voting and Election Day was positive. There were no noted issues with the Early Voting Centers application and it is recommended it be used again in future elections. There was at least one instance of a voter entering in their physical address into the Polling Place Locator application and it resulting in an incorrect polling place and precinct assignment, but that was the result of the address not placing correctly due to the geocoders that were used (ESRI's default and the West Virginia GIS Technical Center's Composite Geocoder). Because a structure or site address was not available in those geocoders, it assigned the address to the street centerline, which was on the boundary of two precincts - an error by a matter of feet. The voter was eventually directed to the correct polling place after reaching out to the City Clerk, so this problem was resolved. This was an instance of proving the importance of having and maintaining accurate addressing data and not reflective of the application itself, so it is recommended to use this again in future elections with any necessary data updates.

The wide use of the Voting Wait Times form indicated an ease of use across all workers who were trained on using it. The low number of Election Day requests sent to the City Clerk could indicate smooth operations of the polls, but it could also indicate users' unwillingness to submit requests through the form. Conversations with the City Clerk post-election seemed to point to the former more than the latter. It is recommended the Election Assistant application, which combined the Voting Wait Times and Election Requests forms in future elections, be used for future elections.

Finally, the Election Results solution was determined to be problematic on multiple fronts. Firstly, the Dashboard that comes with the solution is meant more for partisan elections, which the City of Morgantown does not hold. This intended use made this version of the Dashboard unusable at this scale. Secondly, Morgantown's Municipal Election is different, as all voters vote for each ward's representative on City Council. This made it more practical to break down the City Council contests by each ward, using the precincts instead of using the ward for the voting jurisdictions, which added to the data entry due to there being nine precincts and only seven wards. Finally, the votes were separated into types: Election Day, Early Voting, and Mail-in/Absentee. This bogged down the data entry time due to the need to enter a new attribute table row for every vote type, contest, jurisdiction, and candidate, resulting in 405 entries into the attribute table.

One recommendation is to pre-build the attribute table into a CSV (comma-separated values) file that can be used in Excel, which has all the vote type, contest, jurisdiction, and candidate listed, with only the vote number needing entered. Another possibility is making each vote type a separate attribute in the table, so each row would only need to be broken down by contest, jurisdiction, and candidate, which would result in approximately only 1/3 (33%) of the necessary attribute table rows. The total vote could then be calculated for the "Number of Votes" field. With either option, the recommendation is to avoid, as much as possible, forcing non-GIS staff to manually enter all of this data in ArcGIS Pro on the night of the election and instead complete the results in a program like Excel, then appending the data into the Election Results table using the Append tool in ArcGIS Pro or ArcGIS Online. The failure to do this step beforehand prevented unveiling the Election Results Dashboard until long after the results were known. If this is to be used in future elections, it is recommended going back to the drawing board to have a clear and concise workflow to get the results posted as quickly and efficiently as possible.

The responses in the Voter Satisfaction Survey were overwhelmingly positive which reflects well on the operations of the election. The response rate would also suggest that the distribution of flyers and QR codes, along with posts on social media were successful, but there remains room for improvement. The form seemed to work well as users were able to provide many answers through simply selecting an option instead of typing, and when asked for comments, many respondents did so with some in detail. It is recommended to continue use of this survey in future elections.

While most of the solutions appeared to have been greatly successful, there is room for improvement evident, especially with regards to posting the results of elections into a Dashboard. Ultimately, the use of the ArcGIS Solutions for the Municipal Election proved quite effective as they did not require intensive configuration from what was provided "out of the box." It would be ideal to continue the use of these very same applications, however in two years' time, potential improvements may require redeployment and reconfiguration of new versions. Knowing the workflow of deployment and configuration will cut down on the necessary development time. Whether or not redeployment is necessary should be known by the end of the 2022 calendar year.

The goals for the next election cycle are:

1. Updating any necessary data for the solutions.
2. Improving data entry workflow for Election Results.
3. Improve visibility and usage of City Council, Early Voting Center, and Polling Place applications.
4. Improve the response rate of the Voter Satisfaction Survey.

The embracement of GIS technology in the 2021 Municipal Election proved to be a positive experience for officials, staff, and citizens overall. I believe the improvements to communication and outreach for this election to be a major accomplishment for the City's growing GIS program and I am proud to have been part of the effort. I credit the City Clerk's Office, Christine Wade and Heather Carl, along with all the head commissioners and poll workers who utilized and contributed to the success of this technology throughout the election.