sapor 0.3.6

data/lib/sapor/pseudorandom_multirange_enumerator.rb

@@ -0,0 +1,87 @@
+# encoding: utf-8
+#
+# Statistical Analysis of Polling Results (SAPoR)
+# Copyright (C) 2020 Filip van Laenen <f.a.vanlaenen@ieee.org>
+#
+# This file is part of SAPoR.
+#
+# SAPoR is free software: you can redistribute it and/or modify it under the
+# terms of the GNU General Public License as published by the Free Software
+# Foundation, either version 3 of the License, or (at your option) any later
+# version.
+#
+# SAPoR is distributed in the hope that it will be useful, but WITHOUT ANY
+# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
+# A PARTICULAR PURPOSE.  See the GNU General Public License for more details.
+#
+# You can find a copy of the GNU General Public License in /doc/gpl.txt
+#
+
+require 'prime'
+
+module Sapor
+  #
+  # A pseudo-random multi-range enumerator. It uses prime numbers to iterate
+  # in a pseudo-random way through the hyper-rectangle with dimensions
+  # provided as the constructor parameter. If a suitable set of primes can't
+  # be found, the constructor will raise an exception.
+  #
+  class PseudoRandomMultiRangeEnumerator
+    def initialize(dimensions)
+      @dimensions = dimensions
+      @incrementers = calculate_incrementers
+    end
+
+    def each
+      return enum_for { @dimensions.inject(:*) } unless block_given?
+      counters = Array.new(@dimensions.size, 0)
+      loop do
+        yield(counters.dup)
+        @incrementers.each_with_index do |incrementer, i|
+          counters[i] = (counters[i] + incrementer).modulo(@dimensions[i])
+        end
+        break if all_zeros?(counters)
+      end
+    end
+
+    private
+
+    def calculate_incrementers
+      incrementers = []
+      incrementer = 0
+      calculate_incrementer_indexes.each do |i|
+        incrementer = next_incrementer(incrementer)
+        if @dimensions[i] < incrementer
+          fail(ArgumentError, 'Could not construct suitable incrementers.')
+        end
+        incrementers[i] = incrementer
+      end
+      incrementers
+    end
+
+    def calculate_incrementer_indexes
+      @dimensions.each_with_index.sort.map(&:last)
+    end
+
+    def next_incrementer(incrementer)
+      incrementer += 1
+      while incrementer > 1 &&
+            (!prime?(incrementer) || !relative_prime?(incrementer))
+        incrementer += 1
+      end
+      incrementer
+    end
+
+    def prime?(number)
+      Prime.prime?(number)
+    end
+
+    def relative_prime?(incrementer)
+      @dimensions.map { |dimension| dimension.gcd(incrementer) }.max == 1
+    end
+
+    def all_zeros?(counters)
+      counters.max == 0
+    end
+  end
+end

data/lib/sapor/referendum_polychotomy.rb

@@ -0,0 +1,165 @@
+# encoding: utf-8
+#
+# Statistical Analysis of Polling Results (SAPoR)
+# Copyright (C) 2020 Filip van Laenen <f.a.vanlaenen@ieee.org>
+#
+# This file is part of SAPoR.
+#
+# SAPoR is free software: you can redistribute it and/or modify it under the
+# terms of the GNU General Public License as published by the Free Software
+# Foundation, either version 3 of the License, or (at your option) any later
+# version.
+#
+# SAPoR is distributed in the hope that it will be useful, but WITHOUT ANY
+# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
+# A PARTICULAR PURPOSE.  See the GNU General Public License for more details.
+#
+# You can find a copy of the GNU General Public License in /doc/gpl.txt
+#
+
+module Sapor
+  #
+  # Represents a polychotomy for a referendum.
+  #
+  class ReferendumPolychotomy < Polychotomy
+    include NumberFormatter
+
+    def initialize(results, area, dichotomies, max_error)
+      super(results, area, dichotomies, max_error)
+    end
+
+    def refine
+      no_of_new_simulations = 0
+      new_votes = create_new_votes_distributions
+      while @no_of_data_points == 0 || no_of_new_simulations == 0 ||
+            no_of_new_simulations < @no_of_simulations
+        no_of_new_simulations += try_next_data_point(new_votes)
+      end
+      unless @no_of_simulations == 0
+        @error_estimate = calculate_error_estimate(new_votes)
+      end
+      @votes = merge_distributions(@votes, new_votes)
+      @no_of_simulations += no_of_new_simulations
+    end
+
+    def try_next_data_point(new_votes)
+      data_point = next_data_point
+      if data_point[OTHER] >= 0
+        simulate(new_votes, data_point)
+        new_simulation = 1
+      else
+        new_simulation = 0
+      end
+      @no_of_data_points += 1
+      new_simulation
+    end
+
+    def report
+      choice_lengths = @choices.map(&:length)
+      choice_lengths << 6
+      max_choice_width = choice_lengths.max
+      sorted_choices = sort_choices_by_result
+      choice_lines = sorted_choices.map.with_index do |choice, i|
+        next_choice = sorted_choices[i + 1]
+        create_choice_report_line(choice, i, next_choice, max_choice_width)
+      end
+      'Most probable rounded fractions, fractions and 95% confidence' \
+      " intervals:\n" + 'Choice'.ljust(max_choice_width) +
+        '  Result    MPRF    MPF      CI(95%)      P(>↓)   P(R↑)    P(R)' \
+        "   P(R↓)\n" + choice_lines.join("\n")
+    end
+
+    def progress_report
+      space_size_ratio = space_size.to_f / @no_of_data_points
+      if space_size_ratio > 10
+        space_size_ratio = with_thousands_separator(space_size_ratio.round)
+      else
+        space_size_ratio = space_size_ratio.round(1)
+      end
+      "#{with_thousands_separator(@no_of_simulations)} simulations out of " \
+      "#{with_thousands_separator(@no_of_data_points)} data" \
+      " points, 1 / #{space_size_ratio} of search space size" \
+      " (#{with_thousands_separator(space_size)})."
+    end
+
+    def write_outputs(filename)
+      write_rankings(filename)
+    end
+    
+    private
+
+    def create_new_votes_distributions
+      votes = {}
+      @choices.each do |choice|
+        next if choice == OTHER
+        votes[choice] = CombinationsDistribution.new
+        @ranges[choice].each do |value|
+          votes[choice][value] = 0.to_lf
+        end
+      end
+      votes
+    end
+
+    def simulate(votes, data_point)
+      combinations = 1.to_lf
+      data_point.each do |choice, value|
+        combinations *= BinomialsCache.binomial(value, @results[choice])
+      end
+      @combinations_sum += combinations
+      data_point.each do |choice, value|
+        votes[choice][value] += combinations unless choice == OTHER
+      end
+      @choices.each do |a|
+        @choices.each do |b|
+          if data_point[a] > data_point[b]
+            @comparisons[a + '>' + b] += combinations
+          end
+        end
+      end
+      @choices.sort { |a, b| data_point[b] <=> data_point[a] }.each_with_index do |choice, i|
+        @rankings[choice + '@' + i.to_s] += combinations
+      end
+    end
+
+    def calculate_error_estimate(new_simulations)
+      error_estimate = 0
+      @choices.each do |choice|
+        next if choice == OTHER
+        mpv_new = calculate_most_probable_fraction(choice, new_simulations)
+        mpv_old = calculate_most_probable_fraction(choice, @votes)
+        mpv_delta = (mpv_new - mpv_old).abs
+        mprf_new = calculate_most_probable_rounded_fraction(choice, new_simulations)
+        mprf_old = calculate_most_probable_rounded_fraction(choice, @votes)
+        mprf_delta = (mprf_new - mprf_old).abs
+        error_estimate = [error_estimate, mpv_delta, mprf_delta].max
+      end
+      error_estimate
+    end
+
+    def merge_distributions(distributions1, distributions2)
+      merged_distributions = {}
+      @choices.each do |choice|
+        unless choice == OTHER
+          merged_distributions[choice] = distributions1[choice] + \
+                                         distributions2[choice]
+        end
+      end
+      merged_distributions
+    end
+
+    def create_choice_report_line(choice, i, next_choice, max_choice_width)
+      ci_values = @votes[choice].confidence_interval(0.95)
+      confidence_interval = ci_values.map { |x| x.to_f / @area.population_size }
+      choice.ljust(max_choice_width) + '  ' + \
+        six_char_percentage(result(choice)) + '  ' + \
+        six_char_percentage(most_probable_rounded_fraction(choice)) + '  ' + \
+        six_char_percentage(most_probable_fraction(choice)) + '  ' + \
+        six_char_percentage(confidence_interval.first) + '–' + \
+        six_char_percentage(confidence_interval.last) + '  ' + \
+        (next_choice.nil? ? '      ' : six_char_percentage(larger_than(choice, next_choice))) + '  ' + \
+        (i == 0 ? '      ' : six_char_percentage(ranking(choice, i - 1))) + '  ' + \
+        six_char_percentage(ranking(choice, i)) + \
+        (next_choice.nil? ? '' : '  ' + six_char_percentage(ranking(choice, i + 1)))
+    end
+  end
+end

data/lib/sapor/regional_data/area.rb

@@ -0,0 +1,100 @@
+# encoding: utf-8
+#
+# Statistical Analysis of Polling Results (SAPoR)
+# Copyright (C) 2020 Filip van Laenen <f.a.vanlaenen@ieee.org>
+#
+# This file is part of SAPoR.
+#
+# SAPoR is free software: you can redistribute it and/or modify it under the
+# terms of the GNU General Public License as published by the Free Software
+# Foundation, either version 3 of the License, or (at your option) any later
+# version.
+#
+# SAPoR is distributed in the hope that it will be useful, but WITHOUT ANY
+# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
+# A PARTICULAR PURPOSE.  See the GNU General Public License for more details.
+#
+# You can find a copy of the GNU General Public License in /doc/gpl.txt
+#
+
+module Sapor
+  #
+  # Superclass for any area (country, region, city, etc...) for which
+  # polls can be registered.
+  #
+  class Area
+    def threshold
+      nil
+    end
+
+    def lines_to_election_results(lines, capped = false)
+      results = {}
+      candidates = {}
+      lines.each_line do |line|
+        add_line_to_election_results(results, candidates, line, capped)
+      end
+      [results, candidates]
+    end
+
+    def load_capped_election_results(filename)
+      load_full_election_results(filename, true)
+    end
+
+    def load_election_results(filename)
+      load_full_election_results(filename, false).first
+    end
+
+    def load_full_election_results(filename, capped)
+      dirname = File.dirname(File.expand_path(__FILE__))
+      full_filename = File.expand_path(filename, dirname)
+      lines = File.read(full_filename)
+      lines_to_election_results(lines, capped)
+    end
+
+    def summarize_election_results(election_results)
+      summary = {}
+      election_results.each_value do |local_results|
+        add_local_results_to_summary(summary, local_results)
+      end
+      summary
+    end
+
+    private
+
+    def add_line_to_election_results(results, candidates, line, capped)
+      return if line.strip.empty? || line.strip.start_with?('#')
+      values = line.split('|')
+      fail(ArgumentError, "Broken line: #{line}.") unless values.size.equal?(capped ? 4 : 3)
+      add_line_values_to_election_results(results, candidates, values, capped)
+    end
+
+    def add_line_values_to_election_results(results, candidates, values, capped)
+      constituency = values[0].strip
+      unless results.key?(constituency)
+        results[constituency] = {}
+        candidates[constituency] = {}
+      end
+      choice = values[1].strip
+      votes = values[2].gsub(',', '').to_i
+      if results[constituency].key?(choice)
+        fail(ArgumentError, "Choice #{choice} appeared twice in " \
+                            "constituency #{constituency}.")
+      end
+      results[constituency][choice] = votes
+      if capped
+        cap = values[3].gsub(',', '').to_i
+        candidates[constituency][choice] = cap
+      end
+    end
+
+    def add_local_results_to_summary(summary, local_results)
+      local_results.each_pair do |choice, votes|
+        if summary.key?(choice)
+          summary[choice] += votes
+        else
+          summary[choice] = votes
+        end
+      end
+    end
+  end
+end

data/lib/sapor/regional_data/austria.rb

@@ -0,0 +1,84 @@
+#
+# Statistical Analysis of Polling Results (SAPoR)
+# Copyright (C) 2020 Filip van Laenen <f.a.vanlaenen@ieee.org>
+#
+# This file is part of SAPoR.
+#
+# SAPoR is free software: you can redistribute it and/or modify it under the
+# terms of the GNU General Public License as published by the Free Software
+# Foundation, either version 3 of the License, or (at your option) any later
+# version.
+#
+# SAPoR is distributed in the hope that it will be useful, but WITHOUT ANY
+# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
+# A PARTICULAR PURPOSE.  See the GNU General Public License for more details.
+#
+# You can find a copy of the GNU General Public License in /doc/gpl.txt
+#
+
+module Sapor
+  #
+  # The regional data for Austria.
+  #
+  class Austria < Area
+    include Singleton
+
+    def area_code
+      'AT'
+    end
+
+    def coalitions
+      COALITIONS
+    end
+
+    def no_of_seats
+      NO_OF_SEATS
+    end
+
+    def population_size
+      POPULATION_SIZE
+    end
+
+    def seats(simulation)
+      electoral_system.project(simulation)
+    end
+
+    private
+
+    FPO_PARTY = 'Freiheitliche Partei Österreichs'.freeze
+    GRUENE_PARTY = 'Die Grünen–Die Grüne Alternative'.freeze
+    NEOS_PARTY = 'NEOS–Das Neue Österreich und Liberales Forum'.freeze
+    OVP_PARTY = 'Österreichische Volkspartei'.freeze
+    SPO_PARTY = 'Sozialdemokratische Partei Österreichs'.freeze
+
+    COALITIONS = [[FPO_PARTY, OVP_PARTY],
+                  [FPO_PARTY, SPO_PARTY],
+                  [GRUENE_PARTY, NEOS_PARTY, OVP_PARTY],
+                  [GRUENE_PARTY, NEOS_PARTY, SPO_PARTY],
+                  [GRUENE_PARTY, OVP_PARTY],
+                  [GRUENE_PARTY, SPO_PARTY],
+                  [NEOS_PARTY, OVP_PARTY],
+                  [OVP_PARTY],
+                  [OVP_PARTY, SPO_PARTY],
+                  [SPO_PARTY]].freeze
+
+    NO_OF_SEATS = 183
+
+    # Voter turnout on 15 October 2017
+    # Source: Web page with the official results of the elections of 15 October
+    # 2017, downloaded on 29 May 2019,
+    # https://en.wikipedia.org/wiki/2017_Austrian_legislative_election
+    POPULATION_SIZE = 5_069_929
+
+    THRESHOLD = 0.04
+
+    def electoral_system
+      if @electoral_system.nil?
+        @electoral_system = SingleDistrictProportional.new(
+          NO_OF_SEATS, DhondtDenominators, THRESHOLD
+        )
+      end
+      @electoral_system
+    end
+  end
+end

data/lib/sapor/regional_data/belgium-brussels-2014.psv

@@ -0,0 +1,46 @@
+# Statistical Analysis of Polling Results (SAPoR)
+# Copyright (C) 2020 Filip van Laenen <f.a.vanlaenen@ieee.org>
+#
+# This file is part of SAPoR.
+#
+# SAPoR is free software: you can redistribute it and/or modify it under the
+# terms of the GNU General Public License as published by the Free Software
+# Foundation, either version 3 of the License, or (at your option) any later
+# version.
+#
+# SAPoR is distributed in the hope that it will be useful, but WITHOUT ANY
+# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
+# A PARTICULAR PURPOSE.  See the GNU General Public License for more details.
+#
+# You can find a copy of the GNU General Public License in /doc/gpl.txt
+#
+
+# Source: Web page with the official results of the elections of 25 May 2014,
+#         downloaded on 3 November 2016,
+# URL: http://verkiezingen2014.belgium.be/nl/cha/results/results_tab_CKR00000.html
+
+# Replaced FDF with DéFI
+# Replaced PTB-GO! with PTB
+
+Bruxelles-Capitale / Brussel-Hoofdstad | PS                | 123,985
+Bruxelles-Capitale / Brussel-Hoofdstad | MR                | 115,038
+Bruxelles-Capitale / Brussel-Hoofdstad | DéFI              |  55,306
+Bruxelles-Capitale / Brussel-Hoofdstad | Ecolo             |  51,133
+Bruxelles-Capitale / Brussel-Hoofdstad | cdH               |  46,441
+Bruxelles-Capitale / Brussel-Hoofdstad | PTB               |  19,135
+Bruxelles-Capitale / Brussel-Hoofdstad | Open Vld          |  13,283
+Bruxelles-Capitale / Brussel-Hoofdstad | N-VA              |  13,223
+Bruxelles-Capitale / Brussel-Hoofdstad | Debout les Belges |  11,112
+Bruxelles-Capitale / Brussel-Hoofdstad | sp.a              |   9,613
+Bruxelles-Capitale / Brussel-Hoofdstad | ISLAM             |   9,387
+Bruxelles-Capitale / Brussel-Hoofdstad | Parti Populaire   |   8,641
+Bruxelles-Capitale / Brussel-Hoofdstad | CD&V              |   8,173
+Bruxelles-Capitale / Brussel-Hoofdstad | Vlaams Belang     |   5,157
+Bruxelles-Capitale / Brussel-Hoofdstad | La Droite         |   2,165
+Bruxelles-Capitale / Brussel-Hoofdstad | NATION            |   1,623
+Bruxelles-Capitale / Brussel-Hoofdstad | GAUCHES COMMUNES  |   1,445
+Bruxelles-Capitale / Brussel-Hoofdstad | Égalitaires!      |     953
+Bruxelles-Capitale / Brussel-Hoofdstad | PARTI LIBERTARIEN |     750
+Bruxelles-Capitale / Brussel-Hoofdstad | MG                |     567
+Bruxelles-Capitale / Brussel-Hoofdstad | Agora Erasmus     |     382
+Bruxelles-Capitale / Brussel-Hoofdstad | LaLutte-DeStrijd  |     213