Unable to plot multiple lines in a single graph-CodePudding

I am having a weird issue, I am trying to plot multiple lines in a single graph but it is only one. I am sharing the screenshot as you can see the close values are different in both. It is not rendering binance graph as it seems to be overridden.

Graph

Update

The code is given below

# All Imports
import ccxt
import pandas as pd
import matplotlib.pyplot as plt
# Connect binance
binance = ccxt.binance()
ftx = ccxt.ftx()
binance_btc_usdt_ohlcv = binance.fetch_ohlcv('BTC/USDT','1d',limit=100)
ftx_btc_usdt_ohlcv = ftx.fetch_ohlcv('BTC/USDT','1d',limit=100)
df_binance = pd.DataFrame(binance_btc_usdt_ohlcv, columns=['ts', 'o', 'h', 'l', 'c', 'v'])
df_ftx = pd.DataFrame(ftx_btc_usdt_ohlcv, columns=['ts', 'o', 'h', 'l', 'c', 'v'])
fig, ax = plt.subplots()
ax.plot(df_binance['ts'], df_binance['v'],label='Binance')
ax.plot(df_ftx['ts'], df_ftx['v'],label='FTX')

plt.legend()
# ax.tick_params(axis='x', colors='red')
plt.show()

CodePudding user response：

Tested in python 3.8.12, pandas 1.3.3, matplotlib 3.4.3

Existing Code

Works without any issues, however, 'Binance' is small compared to 'FTX', which can be resolved with ax.set_yscale('log')

df_binance = pd.DataFrame(binance_btc_usdt_ohlcv, columns=['ts', 'o', 'h', 'l', 'c', 'v'])
df_ftx = pd.DataFrame(ftx_btc_usdt_ohlcv, columns=['ts', 'o', 'h', 'l', 'c', 'v'])
fig, ax = plt.subplots()
ax.plot(df_binance['ts'], df_binance['v'], label='Binance')
ax.plot(df_ftx['ts'], df_ftx['v'], label='FTX')

ax.legend()
ax.set_yscale('log')  # resolve issues of scale with the y-axis values
plt.show()

Without ax.set_yscale('log'), 'Binance' still shows up on the plot

The text code example in the OP used 'v', but the issue was occuring with 'c' (in the screenshot).
- The issue is df_ftx.c and df_binance.c are almost exactly the same, which we can see by using alpha=0.5.

# plot dataframe
ax = df_binance.plot(x='ts', y='c', label='Binance', figsize=(8, 6), logy=True)
p2 = df_ftx.plot(x='ts', y='c', label='FTX', ax=ax, alpha=0.5)

ax.legend(bbox_to_anchor=(1, 1.02), loc='upper left')
plt.show()

Given the scale of the y-axis, the difference is too small to differentiate the two lines.

>>> df_binance.c.sub(df_ftx.c)

0      1.00
1     -2.13
2      0.07
3     -2.44
4     -0.35
5      1.35
6     11.51
7     -6.17
8    -11.91
9     -2.86
10   -13.98
11    -7.40
12    -3.13
13     1.56
14   -15.52
15    -8.63
16     0.83
17    10.44
18     0.82
19    -0.95
20   -12.82
21    -2.54
22   -15.13
23   -14.46
24    -4.63
25   -12.60
26   -10.01
27   -17.00
28    -4.00
29   -16.00
30    -9.49
31    -5.18
32    -3.71
33    23.95
34    -4.71
35    -2.38
36   -11.53
37    -7.13
38   -10.78
39     1.85
40     0.01
41    -9.68
42     7.87
43     9.90
44    -4.65
45     2.83
46     5.91
47    -3.11
48   -14.48
49   -11.36
50    -0.86
51     2.64
52   -22.12
53    -8.10
54    -6.27
55    -3.69
56    -0.86
57     1.91
58     5.69
59     1.24
60    -1.27
61   -12.48
62    -1.59
63    -8.18
64     5.98
65    -6.26
66    -4.25
67    -2.38
68    11.38
69    -9.39
70    -4.74
71    -0.43
72    -9.36
73    -3.10
74    -0.65
75     1.54
76    -2.72
77    -1.90
78    -0.39
79    -9.10
80    -4.99
81    -6.06
82     6.99
83     0.00
84    -8.78
85     2.43
86    -2.28
87   -10.00
88    -9.65
89    -5.07
90    -1.00
91    -0.06
92   -28.58
93    -8.43
94    -8.67
95   -17.16
96    -3.41
97   -12.59
98    -1.85
99     5.99
Name: c, dtype: float64

Updated Code

Convert 'ts' to a datetime dtype with

Using logy=True instead of secondary_y=True

# plot dataframe
ax = df_binance.plot(x='ts', y='v', label='Binance', figsize=(8, 6), logy=True)
p2 = df_ftx.plot(x='ts', y='v', label='FTX', ax=ax)

ax.legend(bbox_to_anchor=(1, 1.02), loc='upper left')
plt.show()