散景 multi_line 图未显示

Question

我正在尝试在 Bokeh 中绘制 multi_line，但由于某种原因它无法正常工作。我读入了 pandas:

的文件

roads = pd.read_csv('path/file.csv')

生成的数据框如下所示：

    osm_id  code    fclass           name        ref    oneway  maxspeed    layer   bridge  tunnel               x                                              y
0   4264947 5152    cycleway         NaN         NaN      B        0         0        F        F    [4.9223553, 4.9223667, 4.9223853]                   [52.3654363, 52.3654277, 52.3654169]
1   4270144 5124    pedestrian      Dapperplein  NaN      B        0         0        F        F    [4.9279354, 4.9280225, 4.9280546, 4.928041, 4....   [52.3619809, 52.362025, 52.3621167, 52.3622213...
2   4270151 5153    footway      Zeeburgerstraat NaN      B        0         0        F        F    [4.9261096, 4.9260732, 4.9258198]                   [52.3663263, 52.3662732, 52.3661229]
3   4270806 5152    cycleway      Roomtuintjes   NaN      B        0         0        F        F    [4.9308652, 4.9308238, 4.9307548, 4.9304676, 4...   [52.3659402, 52.3657291, 52.3655389, 52.364919...
4   4277440 5124    pedestrian    Vondelpark     NaN      B        0         0        F        F    [4.881478, 4.8816828, 4.881693]                     [52.361566, 52.361699, 52.3617067]

当我尝试用散景绘制坐标时

source = ColumnDataSource(roads)

q = figure(title= "Roads", plot_width=900, plot_height=900)

q.multi_line('x', 'y', source = source, color='blue', line_width=0.5)

show(q)

它输出一个空图。

出了什么问题，有人可以帮助我吗？

谢谢！

Answer 1

问题出在列中的列表，这里是列表的字符串 repr。所以首先将两列都转换为列表::

import ast

roads = pd.read_csv('path/file.csv')

roads['x'] = roads['x'].apply(ast.literal_eval)
roads['y'] = roads['y'].apply(ast.literal_eval)

另一个解决方案：

import ast

conv = lambda x: ast.literal_eval(x)
roads = pd.read_csv('path/file.csv', converters={'x':conv, 'y':conv})

source = ColumnDataSource(roads)

q = figure(title= "Roads", plot_width=900, plot_height=900)

q.multi_line('x', 'y', source = source, color='blue', line_width=0.5)

show(q)

散景 multi_line 图未显示

Bokeh multi_line plot not showing

python

plot

pandas

bokeh