[EL] Election result consolidation system ready for prime time

[Stephanie F Singer]

I should add that, even with documentation, there is definitely a learning curve to using the tool. Please don’t hesitate to contact me directly for “tech support”. 

> On Jan 13, 2022, at 11:01 AM, Stephanie F Singer <sfsinger at campaignscientific.com> wrote:
> 
> 
> With funding from the National Science Foundation (#1936809, #2027089) I just spent two years building a computer system to make it easier to combine and analyze election results from all over the country, despite the huge variety of publication formats chosen by the states, territories and district. 
> 
> The system is open source and well-documented <https://github.com/ElectionDataAnalysis/electiondata>. It can be used to analyze results over time, or to quickly consolidate results as they are published in bulk. So this is not competition for the AP, but it can be used to do sophisticated analysis within hours or a day of publication. It can be used for any granularity, down to precinct or even machine level, if you can feed in granular data. You can play around with one of its visualizations at VerifiedVoting <https://verifiedvoting.org/votevisualizer/#mode/compare>.
> 
> The system has other features, currently more “beta” — i.e., less user-friendly, less well documented —  to merge in census or other data.
> 
> Happy to talk in more detail with anyone who is interested.
> 
> —Stephanie
> 
> electiondata: a Python package for consolidating, checking, analyzing, visualizing and exporting election results
> 
> Here is a short summary from the article in the Journal of Open Source Software <https://joss.theoj.org/papers/10.21105/joss.03739>:
> The software package includes:
> 
> • Process for munging election results from a large variety of files into a single consolidated database.
> 
> –  File types include csv, excel, json, xml (but not pdf), with arbitrary internal struc- ture choices (e.g., xml tags, or column, row, blank line and header choices for flat files). Users do not need to know Python (other than the basics for installing and calling the package).
> 
> –  The system provides detailed messaging and error handling to support the user creating the parameters for a new file format or jurisdiction.
> 
> • Detailed jurisdiction-specific information for all 56 major United States jurisdictions and munging parameters sufficient to process county-level election results from the raw files published by the 56 Boards of Election. Except for the few jurisdictions where only pdf or html files are available, this processing is entirely automatic.
> 
> • Testing of election results in database against reference contest totals.
> • Exports to json and xml NIST Common Data Formats V2 (Wack, 2019), as well as
> 
> exports to tab-separated flat text file.
> • Scatter plot functionality by major subdivision (typically county) for comparing various
> 
> vote counts and census or other external data. See for example Figure 1.
> 
> Figure 1: Sample scatter plot comparing absentee ballot counts for two candidates in different contests.
> 
> <page1image1780822976.png> <page1image1780823264.png> <page1image1780823568.png> <page1image1780823920.png>
> • Algorithmic curation of interesting one-county outliers within contest types (e.g., for all congressional contests in a particular jurisdiction). The algorithm takes into account the size of the outlier relative to the size of the contest margin. See for example Figure 2.
> 
> <page2image1288032832.png>
> 
> Figure 2: Outlier found by algorithm for congressional contests in North Carolina in 2018.
> 
> 
> • Difference-in-difference analysis (following (Herron, 2019)) for contest types where vote counts by type (e.g., election-day, absentee, provisional) are available. 
> 
> _______________________________________________
> Law-election mailing list
> Law-election at department-lists.uci.edu
> https://department-lists.uci.edu/mailman/listinfo/law-election

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://webshare.law.ucla.edu/Listservs/law-election/attachments/20220113/06cbb7fc/attachment.html>