Distances in VAAs

README.md

Charts of Distributions of Candidates or Parties in VAAs

The charts prepared for the article "Measures of Differentiation among Candidates as an Indicator of the Quality of Voting Advice Applications": https://docs.google.com/document/d/1KVTso9cdwvCxhPC3bEp8rAEY_7v3gWvCzHNYVOjcMR8/edit

Includes Wahl-O-Mat (Germany) 2005, 2009, 2013; StemWijzer (Netherlands) 2006, 2010, 2012; and Volební kalkulačka (Czech Republic) 2013

calc_matches_regions.py

'''calculate distances'''

import csv
#from random import shuffle
import copy

path = "/home/michal/project/datapackages/"

#parties
def option2numeric_party(v):
    if v == 'yes':
        return 1
    if v == 'no':
        return -1
    if v == 'abstain':
        return 0
    return 0

# calculate match between two parties u and p
def pmatch(p,q,keys):
    w = 0
    m = 0
    for k in keys:
        try:
            m = m + p[k]*q[k]
            w = w + 1
        except:
            nothing = 0
    if w == 0:
        return 0
    else:
        return m/w   

totals = []

with open("regions.csv") as fin:
    csvr = csv.reader(fin)
    for row in csvr:
        region = row[0]
        region_name = row[1]

        # votes
        party_votes = {}
        with open(path + "vaa-cz-" + region + "-2012-parties-answers/data/votes.csv") as fin:
            csvr = csv.DictReader(fin)
            for row in csvr:
                    for item in row:
                        if option2numeric_party(row['option']) != 'none':
                            try:
                                party_votes[row['voter_id']]
                            except:
                                party_votes[row['voter_id']] = {}
                            party_votes[row['voter_id']][row['vote_event_id']] = option2numeric_party(row['option'])

        #parties
        parties = {}
        with open(path + "vaa-cz-" + region + "-2012-parties-answers/data/voters.csv") as fin:
            csvr = csv.DictReader(fin)
            coverage = 0
            for row in csvr:
                parties[row['id']] = row
                parties[row['id']]['result'] = float(row['result'])/100
                coverage += parties[row['id']]['result']
                

        #get all vote
        vote_events = []
        vedic = {}
        for pkey in party_votes:
            for vekey in party_votes[pkey]:
                vedic[vekey] = vekey

        for key in vedic:
            vote_events.append(key)

    #    raise(Exception)

        print("votes read")       
    #calc match
    #distances = {}

     
        #users x parties
    #i = 0
    #for ukey in user_votes:
    #    for pkey in party_votes:
    #        try:
    #            distances[ukey]
    #        except:
    #            distances[ukey] = {}
    #        distances[ukey][pkey] = match(user_votes[ukey],party_votes[pkey])
    #        i += 1
    #        if i % 1000 == 0:
    #            print(i)


    #csvw = csv.writer(fout)
        
            
        #parties x parties
        #triangle values:
        pdistances_li = []
        pdistances_ind = []
        pdistances_w = []
        pmatches = {}
        ws = []
        wss = []
        for key in party_votes:
            for pkey in party_votes:
                try:
                    pmatches[key]
                except:
                    pmatches[key] = {}
                pmatches[key][pkey] = pmatch(party_votes[key],party_votes[pkey],vote_events)
        pdistances_ind = copy.deepcopy(pmatches)
        for key in party_votes:
            for pkey in party_votes:
                if key == pkey:
                    pdistances_ind[key][pkey] = None
                else:
                    if pdistances_ind[key][pkey] is not None:
                        distance = 1-(pmatches[key][pkey]+1)/2
                        pdistances_li.append({
                            'distance': distance,
                            "c_1": parties[key],
                            'c_2': parties[pkey]
                        })
                        ws.append(1)
                        wss.append([parties[key],parties[pkey]])
                        pdistances_w.append(distance)
                        pdistances_ind[key][pkey] = None
                        pdistances_ind[pkey][key] = None
                
        dists = []
        average = sum(pdistances_w) / sum(ws)
        i = 0
        for item in pdistances_li:
            dists.append(abs(item['distance']-average) * ws[i])
            i += 1
        deviation = sum(dists)/sum(ws)
        print(average, deviation)
        totals.append({
            'average': average,
            'deviation': deviation,
            'id': region,
            'coverage': coverage
        })
    #    raise(Exception)
        
        with open(region + "_distances.csv","w") as fout:
            csvw = csv.writer(fout)
            row = ['distance','id_1','id_2','abbreviation_1','abbreviation_2','result_1','result_2', 'name_1','name_2']
            csvw.writerow(row)
            for r in pdistances_li:
                row = [r['distance'],r['c_1']['id'],r['c_2']['id'],r['c_1']['abbreviation'],r['c_2']['abbreviation'],r['c_1']['result'],r['c_2']['result'],r['c_1']['name'],r['c_2']['name']]
                csvw.writerow(row)

with open("regions_statistics.csv","w") as fout:
    fieldnames = ['id','average','deviation','coverage']
    csvw = csv.DictWriter(fout,fieldnames)
    csvw.writeheader()
    for item in totals:
        csvw.writerow(item)
    

calc_matches_regions2.py

'''calculate distances'''

import csv
#from random import shuffle
import copy

path = "/home/michal/project/datapackages/"

#parties
def option2numeric_party(v):
    if v == 'yes':
        return 1
    if v == 'no':
        return -1
    if v == 'abstain':
        return 0
    return 0

# calculate match between two parties u and p
def pmatch(p,q,keys):
    w = 0
    m = 0
    for k in keys:
        try:
            m = m + p[k]*q[k]
            w = w + 1
        except:
            nothing = 0
    if w == 0:
        return 0
    else:
        return m/w   

totals = []

with open("calcs.csv") as fin:
    csvr = csv.reader(fin)
    for row in csvr:
        calc = row[0]
#        region_name = row[1]

        # votes
        party_votes = {}
        with open(path + calc + "/data/votes.csv") as fin:
            csvr = csv.DictReader(fin)
            for row in csvr:
                    for item in row:
                        if option2numeric_party(row['option']) != 'none':
                            try:
                                party_votes[row['voter_id']]
                            except:
                                party_votes[row['voter_id']] = {}
                            party_votes[row['voter_id']][row['vote_event_id']] = option2numeric_party(row['option'])

        #parties
        parties = {}
        with open(path + calc + "/data/voters.csv") as fin:
            csvr = csv.DictReader(fin)
            coverage = 0
            for row in csvr:
                parties[row['id']] = row
                parties[row['id']]['result'] = float(row['result'])/100
                coverage += parties[row['id']]['result']
                

        #get all vote
        vote_events = []
        vedic = {}
        for pkey in party_votes:
            for vekey in party_votes[pkey]:
                vedic[vekey] = vekey

        for key in vedic:
            vote_events.append(key)

    #    raise(Exception)

        print("votes read")       
    #calc match
    #distances = {}

     
        #users x parties
    #i = 0
    #for ukey in user_votes:
    #    for pkey in party_votes:
    #        try:
    #            distances[ukey]
    #        except:
    #            distances[ukey] = {}
    #        distances[ukey][pkey] = match(user_votes[ukey],party_votes[pkey])
    #        i += 1
    #        if i % 1000 == 0:
    #            print(i)


    #csvw = csv.writer(fout)
        
            
        #parties x parties
        #triangle values:
        pdistances_li = []
        pdistances_ind = []
        pdistances_w = []
        pmatches = {}
        ws = []
        wss = []
        for key in party_votes:
            for pkey in party_votes:
                try:
                    pmatches[key]
                except:
                    pmatches[key] = {}
                pmatches[key][pkey] = pmatch(party_votes[key],party_votes[pkey],vote_events)
        pdistances_ind = copy.deepcopy(pmatches)
        for key in party_votes:
            for pkey in party_votes:
                if key == pkey:
                    pdistances_ind[key][pkey] = None
                else:
                    if pdistances_ind[key][pkey] is not None:
                        distance = 1-(pmatches[key][pkey]+1)/2
                        pdistances_li.append({
                            'distance': distance,
                            "c_1": parties[key],
                            'c_2': parties[pkey]
                        })
                        ws.append(parties[key]['result']*parties[pkey]['result'])
                        wss.append([parties[key],parties[pkey]])
                        pdistances_w.append(distance*parties[key]['result']*parties[pkey]['result'])
                        pdistances_ind[key][pkey] = None
                        pdistances_ind[pkey][key] = None
                
        dists = []
        average = sum(pdistances_w) / sum(ws)
        i = 0
        for item in pdistances_li:
            dists.append(abs(item['distance']-average) * ws[i])
            i += 1
        deviation = sum(dists)/sum(ws)
        print(average, deviation)
        totals.append({
            'average': average,
            'deviation': deviation,
            'id': calc,
            'coverage': coverage
        })
    #    raise(Exception)
        
        with open(calc + "_distances.csv","w") as fout:
            csvw = csv.writer(fout)
            row = ['distance','id_1','id_2','abbreviation_1','abbreviation_2','result_1','result_2', 'name_1','name_2','color_1','color_2']
            csvw.writerow(row)
            for r in pdistances_li:
                row = [r['distance'],r['c_1']['id'],r['c_2']['id'],r['c_1']['abbreviation'],r['c_2']['abbreviation'],r['c_1']['result'],r['c_2']['result'],r['c_1']['name'],r['c_2']['name'],r['c_1']['color'],r['c_2']['color']]
                csvw.writerow(row)

with open("calcs_statistics.csv","w") as fout:
    fieldnames = ['id','average','deviation','coverage']
    csvw = csv.DictWriter(fout,fieldnames)
    csvw.writeheader()
    for item in totals:
        csvw.writerow(item)
    

calcs.csv

vaa-de-bundestag-2005-parties-answers
vaa-de-bundestag-2009-parties-answers
vaa-de-bundestag-2013-parties-answers
vaa-nl-tweede-kamer-2006-parties-answers
vaa-nl-tweede-kamer-2010-parties-answers
vaa-nl-tweede-kamer-2012-parties-answers
vaa-cz-psp-2013-parties-answers

calcs_statistics.csv

d3.boxesplot.js

/* requires D3 + https://github.com/Caged/d3-tip */
d3.boxesplot = function() {
    //defaults
    margin = { top: 10, right: 30, bottom: 30, left: 40 };
    function boxesplot(selection) {
        selection.each(function(d, i) {
            //options
            var data = (typeof(data) === "function" ? data(d) : d.data),
                lines = (typeof(lines) === "function" ? lines(d) : d.lines),
                size = (typeof(size) === "function" ? size(d) : d.size);
            
            // chart sizes
            var width = size['width'] - margin.left - margin.right,
                height = size['height'] - margin.top - margin.bottom;
            
            //set up scales
            var x = d3.scale.linear()
			             .domain([-0.09, 1.09])
			             .range([0, width])
			var y = d3.scale.linear()
			             .domain([-0.1, 1.1])
			             .range([height, 0]);
			//set up axes
			var xAxis = d3.svg.axis()
                .scale(x)
                .orient("bottom");
            var yAxis = d3.svg.axis()
                .scale(y)
                .orient("left");
                
            var element = d3.select(this); 
            
            //add x axis
            element.append("g")
              .attr("class", "x axis")
              .attr("transform", "translate(0," + height + ")")
              .call(xAxis)
             .append("text")
              .attr("x", x(1))
              .attr("dx", "1em")
              .attr("dy", "-1em")
              .style("text-anchor", "end")
              .text("Distance");
           
            
            var line = d3.svg.line()
                .x(
                    function(d) { 
                        return d[0]; })
                .y(function(d) { return d[1]; });

            
            centerline = Array([x(lines["mean"]),y(-0.1)],[x(lines["mean"]),y(1.1)]);
            
            element.append("path")
                  .datum(centerline)
                  .attr("class", "centerline")
                  .attr("d", line);
            
            deviationlines = [
                [
                    [x(lines["mean"]+lines["dev"]),y(-0.1)],
                    [x(lines["mean"]+lines["dev"]),y(1.1)]
                ],
                [
                    [x(lines["mean"]-lines["dev"]),y(-0.1)],
                    [x(lines["mean"]-lines["dev"]),y(1.1)]
                ],
            ];
            
            element.selectAll(".deviationline")
                .data(deviationlines)
                    .enter()
                .append("path")
                  .datum(function(d) {
                    return d;
                  })
                  .attr("class", "deviationline")
                  .attr("d", line);
            
            coef = .5;
            
            //add randoms
            for (i in data) {
                data[i]['random'] = Math.random();
            }
            
            //create boxes
            element.selectAll(".rectangle")
                .data(data)
                    .enter()
                .append("rect")
                    .attr("x",  function (d) {
                        a = Math.max(d.result_1,d.result_2);
                        return x(d.distance-coef*a/2)
                     })
                    .attr("y", function (d) {
                        b = Math.min(d.result_1,d.result_2);
                        
                        return y(d.random*0.75+0.25-coef*b/2);
                     })
                	.attr("width", function (d) {
                	    a = Math.max(d.result_1,d.result_2);
                	    return x(coef*a)-x(0);
                	})
                	.attr("height", function (d) {
                	    b = Math.min(d.result_1,d.result_2);
                	    return -1*y(7/2*coef*b)+y(0);
                	}) //7/2 height vs. width of chart
                	//.attr("stroke-width", function(d) {return d.r2})
                	.attr("class", function(d) {return d.region})
                	.attr("title", function(d) {return d.distance + ":" + d.result_1 + ":" + d.result_2});
            
            //create lines around boxes - longer
            element.selectAll(".lines")
                .data(data)
                    .enter()
                .append("path")
                    .attr("class","line")
                    .attr("stroke",function(d) {
                        if ((d.color_1 != '') && (d.color_2 != '')) {
                            if (d.result_1>d.result_2)
                                return d.color_1;
                            else
                                return d.color_2;
                        }
                    })
                    .attr("d",function(d) {
                        a = Math.max(d.result_1,d.result_2);
                        b = Math.min(d.result_1,d.result_2);
                        xx = x(d.distance-coef*a/2);
                        yy = y(d.random*0.75+0.25-coef*b/2);
                        width = x(coef*a)-x(0);
                        l = Array([xx,yy],[xx+width,yy]);
                        return line(l);
                    })
            //create lines around boxes - longer 2
            element.selectAll(".lines")
                .data(data)
                    .enter()
                .append("path")
                    .attr("class","line")
                    .attr("stroke",function(d) {
                        if ((d.color_1 != '') && (d.color_2 != '')) {
                            if (d.result_1>d.result_2)
                                return d.color_1;
                            else
                                return d.color_2;
                        }
                    })
                    .attr("d",function(d) {
                        a = Math.max(d.result_1,d.result_2);
                        b = Math.min(d.result_1,d.result_2);
                        xx = x(d.distance-coef*a/2);
                        yy = y(d.random*0.75+0.25-coef*b/2)-1*y(7/2*coef*b)+y(0);
                        width = x(coef*a)-x(0);
                        l = Array([xx,yy],[xx+width,yy]);
                        return line(l);
                    })
            //create lines around boxes - shorter
            element.selectAll(".lines")
                .data(data)
                    .enter()
                .append("path")
                    .attr("class","line")
                    .attr("stroke",function(d) {
                        if ((d.color_1 != '') && (d.color_2 != '')) {
                            if (d.result_1>d.result_2)
                                return d.color_2;
                            else
                                return d.color_1;
                        }
                    })
                    .attr("d",function(d) {
                        a = Math.max(d.result_1,d.result_2);
                        b = Math.min(d.result_1,d.result_2);
                        xx = x(d.distance-coef*a/2);
                        yy = y(d.random*0.75+0.25-coef*b/2);
                        height = -1*y(7/2*coef*b)+y(0);
                        l = Array([xx,yy],[xx,yy+height]);
                        return line(l);
                    })
            //create lines around boxes - shorter 2
            element.selectAll(".lines")
                .data(data)
                    .enter()
                .append("path")
                    .attr("class","line")
                    .attr("stroke",function(d) {
                        if ((d.color_1 != '') && (d.color_2 != '')) {
                            if (d.result_1>d.result_2)
                                return d.color_2;
                            else
                                return d.color_1;
                        }
                    })
                    .attr("d",function(d) {
                        a = Math.max(d.result_1,d.result_2);
                        b = Math.min(d.result_1,d.result_2);
                        xx = x(d.distance-coef*a/2)+x(coef*a)-x(0);
                        yy = y(d.random*0.75+0.25-coef*b/2);
                        height = -1*y(7/2*coef*b)+y(0);
                        l = Array([xx,yy],[xx,yy+height]);
                        return line(l);
                    })
                	
        });
    }
    //
    boxesplot.data = function(value) {
        if (!arguments.length) return value;
            data = value;
        return boxesplot;
    }
    boxesplot.lines = function(value) {
        if (!arguments.length) return value;
            lines = value;
        return boxesplot;
    }
    boxesplot.size = function(value) {
        if (!arguments.length) return value;
            size = value;
        return boxesplot;
    }
    return boxesplot;
}

index.html

<!DOCTYPE html>
<html lang="en">
  <head>
    <meta charset="utf-8">
    <title>Chart</title>
	<meta name="viewport" content="width=device-width, initial-scale=1.0">
    <meta name="description" content="plot">
    <meta name="author" content="Michal Škop">
    <link rel="stylesheet" href="//cdn.bootcss.com/bootswatch/3.3.5/journal/bootstrap.min.css">
	
    <script src="//cdn.bootcss.com/d3/3.5.6/d3.min.js"></script>
    <script src="//cdn.bootcss.com/jquery/1.11.3/jquery.min.js"></script>
<!--    <script src="./modernizr.svg.js"></script>-->
    <script src="//cdn.bootcss.com/modernizr/2.8.3/modernizr.min.js"></script>
    <script src="./d3.boxesplot.js"></script>
    

     
     <style type="text/css">
		text {
		  font-family: sans-serif;
		}
		
		.tick {
		  fill-opacity: 0;
		  stroke: #000000;
		  stroke-width: 1;
		}
		
		.domain {
		    fill: none;
			fill-opacity: 0;
			stroke: black;
			stroke-width: 1;
		}
		.axis line {
			fill: none;
			fill-opacity: 0;
			stroke: black;
			stroke-width: 1;
			shape-rendering: crispEdges;
		}
		
		.axis text {
			font-family: sans-serif;
			font-size: 15px;
		}
		.axis {
		  
		}
		rect, circle {
		  fill-opacity: .15;
		  stroke-opacity: 0.99;
		  stroke-width: 1;
		  fill: blue;
		  
		}

		.line {
		  /*stroke: #000;*/
		  stroke-width: 2;
		  stroke-opacity: 1;
		}
		.centerline {
		  stroke: #888;
		  stroke-width: 3;
		  stroke-opacity: 1;
		}
		.deviationline {
		    stroke: #888;
		    stroke-width: 2;
		    stroke-opacity: 1;
		    stroke-dasharray: 5,5;
		}

		.label {
		  font-family: sans-serif;
		  font-size: 15px;
		}
		.center, .madari {
		  fill: #b00;
		  stroke: #b00;
		}
        .estate, .banska-bystrica, .brezno {
		  fill: #080;
		  stroke: #080;
		}
		
	</style>
  </head>
  <body>
  <nav class="navbar navbar-default" role=navigation">
      <div class="container">
        <div class="navbar-header">
          <a class="navbar-brand">Distances of parties in Czech VAA 2013</a>
        </div>
      </div>
  </nav>
  <div class="alert alert-info">Every rectangle represents one pair of parties. The <strong>size</strong> of rectangles represents number of voters for the 2 parties.</div>
  <div id="chart"></div>
<script>

var $chart = $('#chart');

function drawGraphic() {
 // clear out existing graphics
    $chart.empty();
    var margin = { top: 10, right: 30, bottom: 30, left: 40 };
    var width = $chart.width() - margin.left - margin.right;
    var boxesplot = [{
        "data": xdata,
        "size": {
            "width": width,
            "height": 2/7*width
        },
        "lines": {"mean":0.3715520140531649,"dev":0.13}
    }];
    var svg = d3.select("#chart")
            .append("svg")
            .attr("width",boxesplot[0]['size']['width'])
            .attr("height",boxesplot[0]['size']['height']);
    
    var bp = d3.boxesplot()
            .data(function(d) {return d.data})
            .size(function(d) {return d.size})
            .lines(function(d) {return d.lines});

            
    svg.selectAll(".boxesplot")
        .data(boxesplot)
      .enter()
        .append("svg:g")
        .attr("transform", "translate(" + margin.left + "," + margin.top + ")")
        .call(bp);
}

if (Modernizr.svg) { // if svg is supported, draw dynamic chart
    d3.csv("vaa-cz-psp-2013-parties-answers_distances_selected.csv", function(data) {
        xdata = data;
        drawGraphic();
        window.onresize = drawGraphic;
    })
}   
 </script>
  </body>
</html>    

jihocesky-kraj_distances.csv

jihomoravsky-kraj_distances.csv

karlovarsky-kraj_distances.csv

kralovehradecky-kraj_distances.csv

liberecky-kraj_distances.csv

moravskoslezsky-kraj_distances.csv

olomoucky-kraj_distances.csv

pardubicky-kraj_distances.csv

plzensky-kraj_distances.csv

regions.csv

regions_statistics.csv

stredocesky-kraj_distances.csv

thumbnail.png

ustecky-kraj_distances.csv

vaa-cz-psp-2013-parties-answers_distances.csv

vaa-cz-psp-2013-parties-answers_distances_selected.csv

vaa-de-bundestag-2005-parties-answers_distances.csv

vaa-de-bundestag-2009-parties-answers_distances.csv

vaa-de-bundestag-2013-parties-answers_distances.csv

vaa-nl-tweede-kamer-2006-parties-answers_distances.csv

vaa-nl-tweede-kamer-2010-parties-answers_distances.csv

vaa-nl-tweede-kamer-2012-parties-answers_distances.csv

vysocina_distances.csv

zlinsky-kraj_distances.csv