holoviews按大小和标记绘制数据框

时间:2019-05-13 11:38:09

标签: python data-visualization holoviews

如何在具有x,y,label_color,label_marker列的数据框中分散绘图数据,以使x,y为坐标,相同的label_color值提供相同的颜色(或label_color为颜色),相同的label_marker提供相同的标记(或label_marker是标记符号)。理想情况下,图例将具有有关此跨产品标签的信息。

我已经尝试过df.hvplot.scatter(x="x", y="y", c="label_color"),但是效果很好。我尝试使用s =“ label_size”代替标记,但是我希望有一个清晰的标记[pan]。这也给人一个奇怪的传说。我进一步尝试了df.hvplot(kind="points", x="1",y="2", c="label_color", s="label_size", by="other_label")

例如

df = pd.DataFrame([[1, 1, "a", "A"], 
                   [2, 2, "a", "B"], 
                   [3,3, "b", "C"],
                   [3,4, "b", "C"]], columns=["x", "y", "color_label", "marker_label"])
scatter_by_color_and_marker(df).redim.range(x=(0,5), y=(0,5))

更新: 我实现了以下内容。欢迎更简单的实现/增强功能

class PianoIter(object):
    def __init__(self, max_iter=100):
        self.max_iter = max_iter
    def __iter__(self):
        self.a = 1
        return self
    def next(self):
        x = self.a
        self.a += 1
        if x>self.max_iter:
            raise Exception("too large iteration")
        return x

from itertools import cycle as Cycle
class DictCycle(dict):
    def __init__(self, d=None, cycle=None):
        d = d if d is not None else {}
        super(DictCycle, self).__init__(d)
        self.fixed_keys = d.keys()
        self.cycle = Cycle(cycle) if cycle is not None else iter(PianoIter())
    def __getitem__(self, label):
        if label not in self:
            value = next(self.cycle)
            self[label] = value
            return value
        else:
            return self.get(label)
def scatter_by_color_and_marker(df, x_label="x", y_label="y", color_label="color_label", marker_label="marker_label", color_dict=None, marker_dict=None):
    color_dict = DictCycle(d=color_dict, cycle=hv.Cycle().values)
    marker_dict = DictCycle(d=marker_dict, cycle=["o", "+", "d", "*", "s", "v", "^", "<", ">", "x"])
    l = []
    for (c_l, m_l), df_i in df.groupby([color_label, marker_label]):
        c = color_dict[c_l]
        m = marker_dict[m_l]
        l.append(df_i.hvplot.scatter(x=x_label, y=y_label).options(color=c, marker=m))
    for c_l, c in color_dict.items():
        l.append(hv.Curve([], label="C "+str(c_l)).opts(line_color=c))
    for m_l, m in marker_dict.items():
        l.append(hv.Points([None,None], label="M "+str(m_l)).opts(marker=m, color="black"))
    return hv.Overlay(l)

0 个答案:

没有答案