Feature Request: stack another indicator on the mpf.plot() with axes.twinx()

[ggg5269]

Is it able to plot a data from pd.DataFrame() (or pd.Series()) object on the axes after mpf.plot()? like plot on to the secondary y axis.

actually the figure i'm going to plot is place a drought index on to corn futures ohlc candlestick plot, somehow it shows up an error with datetime format.

sorry about my grammar, i can post the error message tomorrow and also source data or screenshots.

[DanielGoldfarb]

Did you read the documentation?

Please read, completely, the following, in this order, and then see if maybe your question is answered:

1. https://github.com/matplotlib/mplfinance/blob/master/examples/addplot.ipynb
2. https://github.com/matplotlib/mplfinance/blob/master/examples/panels.ipynb
3. https://github.com/matplotlib/mplfinance/blob/master/markdown/subplots.md

Thanks. All the best. --Daniel

[ggg5269]

Hi Daniel,
thanks for your reply.
I've read all the doc and example code, and make some adjustment of my code.

fig = mpf.figure(figsize=(16, 7), tight_layout=True,style='default')
ax1 = fig.add_subplot(111)
apds = [
    mpf.make_addplot(drought.sum(1), ax=ax1, secondary_y=True),
]

mpf.plot(c,addplot=apds, ax=ax1,volume=False)

this is what i get after setting secondary y axis, but still have the candlestick at the bottom.
where is my mistake?

[DanielGoldfarb]

@ggg5269
In my experience, problems similar to this are usually caused by data (not code), either because of the data itself, or possibly the way the data is formatted. I do not see anything obviously wrong with the code.

(That said, as a general rule I would strongly recommend against creating and using Figure and Axes, as you are, unless you are trying to accomplish something that is otherwise not supported by mplfinance; however it appears to me that what you are trying to do is definitely supported by a simple use of mpf.make_addplot() and mpf.plot() without any need to access the Figure or Axes directly. But, for now, don't bother changing the code, because it should be working the way you have it. First let's address the problem you are seeing.)

So in order to help you solve this problem, I would need to see the data you are using, and the code that processes the data.

Please attach you data (preferably as csv files, or as whatever format you are using), and include also all of the code that prepares that data, reading it in, and placing it into the variables that get passed to mplfinance.

Thank you. --Daniel

[ggg5269]

agri.xlsx
here's how iI read the excel file for corn future price.

import pandas as pd

c = pd.read_excel("price_data/agri.xlsx", sheet_name='C 1 Comdty')
c.columns = ['Date', 'Open', 'High', 'Low', 'Close', 'Volume']
c = c.set_index('Date')

And for the drought index,
dm_export_20000101_20201026.zip

drought_df = pd.read_csv("dm_export_20000101_20201026.csv", parse_dates=['MapDate']).sort_values('MapDate').iloc[:, [0, 3, 4, 5, 6, 7]]
drought_df.columns = ['Date', 'D0', 'D1', 'D2', 'D3', 'D4']
drought_df = drought_df.set_index("Date")

drought_df = drought_df.apply(lambda x: x.str.replace(',','')).astype(float, errors='ignore')

And then I do some manipulation to make sure they got same index range

drought = pd.DataFrame(index=c.index)
drought = pd.concat([drought, drought_df], 1).fillna(method='ffill')

Finally I get the fig which was shown in the previous comment.

[DanielGoldfarb]

And then I do some manipulation to make sure they got same index range

Can you please show this code too? I just want to make sure it is not having some unexpected, accidental side-effect on the data. I want to be able to completely reproduce, on my own machine, the problem you are seeing. Thank you.

NEVER MIND I see it now. Please ignore this comment.

[DanielGoldfarb]

OK. I was able to reproduce and I found the problem.
The problem is indeed that you are passing in your own Axes objects, in which case mplfinance can only assume you want to use the specific Axes objects that you pass, which means secondary_y=True must be ignored (because a secondary_y is actually another Axes object).

I strongly recommend that you do not create your own Figure and Axes.

Instead, keep your code simple, as shown in "Fix Number 1" below.
(However if you really want to create and use your own Axes, it can still be done that way; see "Fix Number 2" below.)

---

To fix the problem, do ONE of the following:

---

Fix Number 1 (preferred)

Instead of:

fig = mpf.figure(figsize=(16, 7), tight_layout=True,style='default')
ax1 = fig.add_subplot(111)
apds = [mpf.make_addplot(drought.sum(1), ax=ax1, secondary_y=True),]
mpf.plot(c,addplot=apds, ax=ax1,volume=False)

Do the following:

apds = mpf.make_addplot(drought.sum(1),secondary_y=True)
mpf.plot(c,addplot=apds,volume=False,figsize=(16, 7),tight_layout=True)

Here is the result:

---

Fix Number 2 (NOT preferred)

Add the line: ax2 = ax1.twinx() and use ax2 in make_addplot(), thus:

fig = mpf.figure(figsize=(16, 7), tight_layout=True,style='default')
ax1 = fig.add_subplot(111)
ax2 = ax1.twinx()
apds = [mpf.make_addplot(drought.sum(1), ax=ax2, secondary_y=True),]
mpf.plot(c,addplot=apds, ax=ax1,volume=False)

Here is the result:

---

NOTICE THAT in "Fix Number 2" you can see a vertical line on the left side at the beginning of the drought data. This is because the first row in the drought data is all NaN. (Therefore, you may want to drop the first row from both data sets). You can see this more clearly if you plot only the first 150 rows (and use plot type='candle') ...

[ggg5269]

wow, thanks for your detailed answer.
now I know what's going on with the axes,
I'll change my code to the first one~

[ChrisChiasson]

This is the most outstanding help from a package maintainer I have read in a very very long time.