Table of Contents

My Georgia campaign finance overview spreadsheets are the gold standard. Democrats and Republicans alike have acknowledged the quality of my work in this area. I take reports filed with the state ethics commission, clean and process the data, and enter basic statistics into my sheets, including breaking out how much each candidate raised from individual contributions, which is not available in the pre-prepared report summaries.

I also make spreadsheets showing campaign finance information in Rhode Island. The quality of the raw data there is significantly higher, making my job easier, but as far as I know, I am still the only person putting it all in one place.

When a federal or state office is on the ballot, Georgia does a great job making the results easily available (with a couple of exceptions). But when that's not the case, it can be difficult to find county and municipal election results, especially official results. Detailed write-in information is also obscure, even in state races. I collect a lot of this information, and I hope to make it available on my website in the near future. (The Secretary of State's office also collects some of this information but does not post it online.)

I have broken down statewide election results by district in several states, even outside of my focus area – for example, I determined that Gretchen Whitmer won a very narrow majority of Michigan's state House districts in 2018 (map | spreadsheet). In Georgia, I even use demographic information to improve the accuracy of my calculations. In the coming days, I will set up a single page for easy public access to all the calculations I have previously posted publicly.

I hope to soon make available my Georgia county and municipal district calculations as well. Unfortunately, while I am confident my statistical methods improve calculation accuracy, they can only do so much, especially in less populous areas. Once I can provide statistically valid margins of error on these calculations, I will begin making them available publicly; in the meantime, feel free to contact me to request details on specific districts.

Yet another way I've applied data to election results is to help me project (or “call”) the winner of a race before both the major networks and news sources and their newer competitors. Sometimes, this involves modeling. These models are tools to help me follow a race. They are necessarily limited, and are not in themselves predictions; rather, they give me the data I need to apply my skills to provide informed coverage, and eventually to make a call.

Most famously, I modeled the 2020 presidential results in Georgia, but I pioneered a related approach in the March 12, 2019 special election runoff for HD 176, in South Georgia. One week later, using a similar model, I was able to project that the much-anticipated Gwinnett MARTA referendum would fail nearly an hour before the full results came in. That November, I built a similar model that helped me call the Savannah mayoral runoff for Van Johnson.

All those models were based on past results – the first round of the election, for both runoffs, and the 2018 PSC D3 runoff for the MARTA referendum – and utilized breakdowns by both precinct and vote type, as well as the numbers of expected absentees and early votes. My 2020 general election model was a bit different. I created it on November 4, after nearly all Election Day and early in-person votes were counted. Instead of using past election results, the model was based on the absentee results so far (and, again, the counts of expected absentees).

I have also often created spreadsheets to track the results of multi-county elections in other states when there is no central source for results or when other central sources are slow to reflect updates from counties. Examples include my 2018 tracking in CA-21, which enabled me to forecast that T.J. Cox would take the lead before the A.P. had even retracted its erroneous call, and my 2020 Democratic primary tracking in three ideologically significant races: TX-28 (Cuellar/Cisneros), IL-03 (Lipinski/Newman), and MO-01 (Clay/Bush).

In TX-28, this enabled me to provide informed coverage amid confusion over which ballots had been counted in each county. In IL-03, it helped me declare Newman the winner over 24 hours before the A.P. did. In MO-01, where the fact that absentee ballots were reported first led DDHQ and various individuals to declare the wrong winner, I was able to announce Bush's victory before the A.P.'s call and before DDHQ switched its call from Clay to Bush. (I owe partial credit there to @StephieTheLefty, who flagged the fact that the call seemed shaky.)

Shapefiles allow users to load geographic information into mapping software. Though obscure, they are extremely important. I independently produced a shapefile for the 2020 general election precincts in Georgia. I have posted this file here in ZIP format. To maintain this file, I monitor voter data and the activities of local boards of elections, among other things. Some of my in-depth documentation is available here. I have not completed the online documentation, in large part because the state has now posted its own shapefile.

I also have shapefiles for many county commission, school board, and city council districts in Georgia. At present, I do not provide these files to others, either for free or at a cost, but I do produce public and private maps based on the files.

I have extracted shapefiles from the Rhode Island reapportionment commission's website, which are posted here. These are not my work except in regards to extraction and formatting.

In Rhode Island, I tracked ideologically significant votes in the General Assembly and calculated ideological scores for legislators based on those votes. This tracker is available here. It has not been updated since midsession.

I'm working on a spreadsheet that shows how every Georgia state legislator voted on every 2021 motion. I already have all the raw vote data entered. It's not suitable for public release, at least not yet – I'm still working on making sure that related votes won't be confused with each other – but keep checking back here to see if that changes. In the meantime, if you are interested in using the spreadsheet for a specific purpose, contact me with an explanation of what you would use it for and I will consider your request.

My master spreadsheet of 2021 elections in Georgia, covering both upcoming special elections and regularly scheduled November elections, can be found here. I add to both lists regularly.

The document doubles as a candidate list. I keep a close eye on candidate filings, and I was the first to report, for example, the mayoral campaign filings of both Felicia Moore and Sharon Gay in Atlanta.

In Georgia, local boards of elections are often governed by obscure, hard-to-find laws, as are municipal elections and some county elections. Municode, a LexisNexis service that a has a near-monopoly on maintaining local codes of ordinances, is inconsistent about providing this information (especially for school boards), and sometimes its information is outright wrong. I am working on compiling the source documents needed to create and maintain comprehensive information on all of this in one place. This is a long-term project, but if you have a question about a specific local government or board of elections, feel free to contact me and I can probably look into it.