Here you will find players' scores for every round ever played, in a form that will be familiar to a wide audience: a Pandas dataframe. This data is updated as soon as possible after the Scorekeeper publishes the 5-round rolling scores report.
This data does not compete with the official record that is published in the Google Group: on the contrary, it is directly derived from it. Here you will find a faithful copy of the official 5-round rolling scores reports, that appear after each round of the game. For rounds when a report was not actually posted, this data will permit you to assemble one, from the scores that appeared prior and subsequent reports. The only difference is that here it is in a form that is more amenable to analysis than tabulations in forum posts.
The rolling scores data is presented in two forms, one that is a close match to the original, row for row with the rolling scores reports, with every round appearing in four successive reports, called raw rolling scores; and another, with a single agreed score per player per round, that is in a form more suitable for data analysis.
“Dixonary Scores for Data Analysis” is a user manual. It explains how the the official rolling scores reports have been interpreted, the layout of the resulting datasets, and how the data for analysis has been derived from the raw data. It also gives practical step-by-step examples of how you can download the data and use it to get answers to interesting questions about the scores that can not be read directly from the available official reports. But it does not try to be a Pandas tutorial. For that, look at the official Pandas site.
“Curating the Official All-time Scores” documents the adjustments that were made in the process of deriving the analysis dataset from the official rolling-scores record.
The 25-Round spreadsheet is an example of what you can do with the dataset provided here. The code to do this is one of the examples in “Dixonary Scores for Data Analysis”.
You are free to download these files for your own purposes. You are not free to redistribute them. If you think someone else should have a copy, send them a link to this page, and not a copy of the file you downloaded, which will start to go out of date very quickly. This is not a burdensome requirement, and its purpose is to inhibit proliferations of scarcely distinguishable versions of the same data. Passing an out-of-date version of these files to someone else is considered modification, since the one you distribute is no longer a true copy of the published version. You will find copyright notices in “Dixonary Scores for Data Analysis”.
The folder Score Archive contains the base data (only periodically updated) for the dataframe presented here.