Option 1. You can plot y1 and y2 on different y axes so that the scale of one does not overwhelm the other one:
fig = plt.figure(figsize=(8,6))
ax1 = fig.add_subplot()
ax2 = ax1.twinx()
ax1.set_ylabel("y1 data")
ax2.set_ylabel("y2 data")
ax1.plot(df.index,df.y1, color = 'red', label = 'y1')
ax2.plot(df.index, df.y2, color = 'blue', label = 'y2')
fig.autofmt_xdate()
ax1.legend()
ax2.legend(loc = 'upper right')
plt.show()
If you have more than two data columns with different orders of magnitude as in your updated question, you have two options:
Option 2. Create multiple subplots. By vertically stacking them, you can re-enforce the fact that they use the same x-axis.
fig,axes = plt.subplots(len(df.columns),1)
for i, ax in enumerate(axes):
ax.set_ylabel(df.columns[i])
ax.plot(df.index,df.iloc[:,i], c = (random.random(),random.random(),random.random()))
fig.autofmt_xdate()
plt.suptitle('Stacked Subplots')
plt.show()
Option 3. You can scale all dataframe columns to the same order of magnitude, plot them is the same plot and include the scale factors in the legend.
def order_of_mag(num):
return math.floor(math.log10(num))
min_order_of_mag = order_of_mag(min(df.min()))
column_scale_factors = [10 ** (order_of_mag(v) - min_order_of_mag) for v in df.min()]
fig,ax = plt.subplots()
for i in range(len(df.columns)):
ax.plot(df.index, df[df.columns[i]]/column_scale_factors[i], label = f'{df.columns[i]} data x {column_scale_factors[i]}')
fig.autofmt_xdate()
plt.xlabel('Date')
plt.legend()
plt.title('Scaled Y Data')
plt.show()
