Question

我的数据中有500个样本（行），存储为csv文件。您可以看到以下5行：

path,Ktype,label,CGX,CGY,C_1,C_2,C_3,C_4,total1,total2,totalI3,total4,feature1,feature2,feature3,feature4,feature5,feature6,feature7,feature8,feature9,feature10,feature11,feature12,A,B,C,D,feature13,feature14,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
    .\Mydata\Case1,k1,1,42,33,0,57.44534,0,52597,71,16,10,276,4038,3789.631,0.6173469,0.6499337,2.103316,0.6661285,1.065539,248.3694,0.630161,0.000192848,0.9999996,0.000642777,1,0,0,1,9.60E-05,3136.698,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
    .\Mydata\Case2,k1,2,163,29,0,43.28862,0,49050,71,16,10,248,2944,2587.956,0.5726808,0.5681185,2.130387,0.601512,1.137578,356.0444,0.6335613,0.000327267,1.000029,0.001271235,1,0,0,1,0.00010854,2676.418,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
    .\Mydata\Case3,k1,3,774,19,0,45.26291,0,53455,71,16,10,212,2366,1982.547,0.408179,0.4579566,1.994296,0.6615351,1.193415,383.4534,0.7153812,0.000264522,1.000031,0.001210507,1,1,0,0,9.54E-05,3221.289,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
    .\Mydata\Case4,k1,4,1116,25,0,80.76469,0,57542,71,16,10,284,3908,3453.988,0.3549117,0.4811547,1.982244,0.6088744,1.131446,454.0122,0.6166388,0.000314288,0.9999836,0.00129846,0,1,1,0,0.000140592,2143.42,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
    .\Mydata\Case5,k1,5,1364,59,1,52.96776,0,49670,71,16,10,228,2725,2642.675,0.4328255,0.475517,1.859871,0.6587288,1.031152,82.32471,0.5775694,0.000466264,0.9999803,0.001765345,0,1,1,0,0.00012014,2439.636,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

我正在为我的数据绘制平行坐标。这是读取csv文件并过滤它的代码的一部分：

d3.csv("Mydata.csv", function(raw_data) {

  // Convert quantitative scales to floats
  data = raw_data.map(function(d) {
    for (var k in d) {
      if (!_.isNaN(raw_data[0][k] - 0) && k != 'id' && k != 'cgX' && k != 'cgY') {
        d[k] = parseFloat(d[k]) || 0;
      }
    };
    return d;
  });

  // Extract the list of numerical dimensions and create a scale for each.
  xscale.domain(dimensions = d3.keys(data[0]).filter(function(k) {
    return (_.isNumber(data[0][k])) && (yscale[k] = d3.scale.linear()
      .domain(d3.extent(data, function(d) { return +d[k]; }))
      .range([h, 0]));
  }));


  // And the rest of the code for drawing parallel coordinates.
  // It is similar to the code in this link:
  //                http://bl.ocks.org/syntagmatic/3150059

}

现在，我想以一种方式改变它，而不是绘制500个样本（平行坐标中有500条折线），它会随机选择100个数据。我该怎么做？

Answer 1

两次通过。

首先阅读所有数据......

d3.csv("Mydata.csv", function(raw_data) {...}

然后随机选择100并将其提供给渲染功能。

奖励是你没有解析/转换你不想渲染的坐标。此外，如果在读取完整500之前达到100，则可以提前退出并跳转到渲染算法。

绘制平行坐标以随机选择数据

1 个答案: