我有一個時間序列的數據幀，其中列是時間值（按順序），每行都是一個單獨的序列。我還有額外的列，提供每行的類別，這反過來又決定了線型和顏色。下面是數據幀：>>> df cat (frac_norm, 2, 1) clustermonth_rel -5 -4 -3 -2 -1 0 1 2 3 4 5 user1 user2 3414845 4232621 -1b 0.760675 0.789854 0.95941 0.867755 0.790102 1 0.588729 0.719073 0.695572 0.647696 0.656323 44369232 3370279 -1b 0.580436 0.546761 0.71343 0.742033 0.802198 0.389957 0.861451 0.651786 0.798265 0.476305 0.896072 022771 3795428 -1b 0.946188 0.499531 0.834885 0.825772 0.754018 0.67823 0.430692 0.353989 0.333761 0.284759 0.260501 22660226 3126314 -1b 0.826701 0.81203 0.765182 0.680162 0.763475 0.802632 1 0.780186 0.844019 0.868698 0.722672 44154510 4348009 -1b 1 0.955656 0.677647 0.911556 0.76613 0.743759 0.61798 0.606536 0.715528 0.614902 0.482267 32860801 164553 -1b 0.870056 0.371981 0.640212 0.835185 0.673108 0.536585 1 0.850242 0.551198 0.873016 0.635556 4120577 3480468 -1b 0.8197 0.879873 0.961178 1 0.855465 0.827824 0.827139 0.304011 0.574978 0.473996 0.358934 3我可以制作下面的圖，其中x軸是 的有序值，顏色取決于 的值，線型取決于 的值。但是，它是逐行的。有沒有辦法對此進行矢量化，例如，通過使用groupby？('frac_norm',2,1)clustercat用于生成圖像的代碼import pandas as pdimport numpy as npcolors = ['r','g','b','c','y','k']lnst = ['-','--']cats = np.sort(df['cat'].unique())clusters = np.sort(df['cluster'].unique())colordict = dict(zip(clusters, colors))lnstdict = dict(zip(cats,lnst))
1 回答
Cats萌萌
TA貢獻1805條經驗 獲得超9個贊
你應該能夠使用熊貓來繪制并避免循環：
from matplotlib.colors import LinearSegmentedColormap
colors = ['r','g','b','c','y','k']
lnst = ['-','--']
cats = np.sort(df['cat'].unique())
clusters = np.sort(df['cluster'].unique())
colordict = dict(zip(clusters, colors))
lnstdict = dict(zip(cats,lnst))
# transpose data frame
df1 = df.T
# map colors from colordict to cluster
cmap = df['cluster'].map(colordict).values.tolist()
# create a custom color map and line style
lscm = LinearSegmentedColormap.from_list('color', cmap)
lstyle = df['cat'].map(lnstdict).values.tolist()
# plot with pandas
df1.iloc[1:12].reset_index(level=0, drop=True).plot(figsize=(20,10),
colormap=lscm,
style=lstyle)
更新（假設您希望兩者都在同一圖表上）
from matplotlib.colors import LinearSegmentedColormap
import matplotlib.pyplot as plt
%matplotlib inline
colors = ['r','g','b','c','y','k']
lnst = ['-','--']
cats = np.sort(df['cat'].unique())
clusters = np.sort(df['cluster'].unique())
colordict = dict(zip(clusters, colors))
lnstdict = dict(zip(cats,lnst))
# transpose data frame
df1 = df.T
# map colors from colordict to cluster
cmap = df['cluster'].map(colordict).values.tolist()
# create a custom color map and line style
lscm = LinearSegmentedColormap.from_list('color', cmap)
lstyle = df['cat'].map(lnstdict).values.tolist()
c = df.columns
# not needed for your actually dataframe
# i am just converting your sample data to numeric
for i in range(len(df.columns[1:])-1):
df[c[i+1]] = pd.to_numeric(df[c[i+1]])
# groupby and get mean of cluster
df2 = df[c[1:]].groupby('cluster').mean()
# create sublots object from matplotlib
fig, ax = plt.subplots()
# add a twin y-axis
ax2 = ax.twiny()
# plot dataframe 1
df1.iloc[1:12].reset_index(level=0, drop=True).plot(ax=ax, figsize=(20,10),
colormap=lscm,
style=lstyle)
# create legend for ax
handles, labels = ax.get_legend_handles_labels()
ax.legend(handles, labels, loc='center left', borderaxespad=-20)
# subplot df2
df2.plot(ax=ax2, colormap='copper')
# create legend for ax2
handles, labels = ax2.get_legend_handles_labels()
ax2.legend(handles, labels, loc='center right', borderaxespad=-20)
添加回答
舉報
0/150
提交
取消