Modified the prior's year script to be more automated. As a result, one should be able to copy the results from the votes details and paste them in the top box of the Jupiter script. The list of votes just need a bit of massage, to add the company as first row entry, and then add " delimiters for strings and the appropriate commas and brackets for the array notation.
Then, the commands can be simply run regardless of the number of companies and/or candidates.
For peace of mind, I ran the same script on last year's election results and it resulted mostly in the same results, with one minor difference. Last year's manual script "randomly" assigned a priority to candidates that had the same score when "summarizing" the votes per company. The new script assign the same "preference" to two or more candidates that have the same score. So if candidate A, B, C, and D got a preference score of 10, 5, 5, and 1, this year's ballot will assign (A=1, B=2, C=2, D=3) whereas the old script appears to have selected one of B or C to be 2nd and 3rd, respectively. It did not change anything to the final ranking.
Please feel free to verify for yourselves the results. Code and all data is here. I included also the raw csv file from the voting site so that you may also verify that there was no error in copying the data over to python.
Modified the prior's year script to be more automated. As a result, one should be able to copy the results from the votes details and paste them in the top box of the Jupiter script. The list of votes just need a bit of massage, to add the company as first row entry, and then add
"delimiters for strings and the appropriate commas and brackets for the array notation.Then, the commands can be simply run regardless of the number of companies and/or candidates.
For peace of mind, I ran the same script on last year's election results and it resulted mostly in the same results, with one minor difference. Last year's manual script "randomly" assigned a priority to candidates that had the same score when "summarizing" the votes per company. The new script assign the same "preference" to two or more candidates that have the same score. So if candidate A, B, C, and D got a preference score of 10, 5, 5, and 1, this year's ballot will assign (A=1, B=2, C=2, D=3) whereas the old script appears to have selected one of B or C to be 2nd and 3rd, respectively. It did not change anything to the final ranking.
Please feel free to verify for yourselves the results. Code and all data is here. I included also the raw csv file from the voting site so that you may also verify that there was no error in copying the data over to python.
civs_ballots.csv