Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

[BUG] "overlay-correlate" example not working

See original GitHub issue

I realize finplot is not a web lib. (Hint: it’s native!)
I’ve read the snippets and not found what I’m looking for.
I’ve searched for any related issues and avoided creating a duplicate issue.
I’ve updated finplot (pip install -U finplot).
I’ve supplied the required data to run my code below.

Code to reproduce

#!/usr/bin/env python3

from datetime import date, timedelta
import finplot as fplt
import pandas as pd
import scipy.optimize
import yfinance as yf

now = date.today()
start_day = now - timedelta(days=55)
df = yf.download('GOOG', start_day.isoformat(), now.isoformat(), interval='90m')
dfms = yf.download('MSFT', start_day.isoformat(), now.isoformat(), interval='90m')

# resample to daily candles, i.e. five 90-minute candles per business day
dfd = df.Open.resample('D').first().to_frame()
dfd['Close'] = df.Close.resample('D').last()
dfd['High'] = df.High.resample('D').max()
dfd['Low'] = df.Low.resample('D').min()

ax,ax2 = fplt.create_plot('Alphabet Inc.', rows=2, maximize=False)
ax2.disable_x_index() # second plot is not timebased

# plot down-sampled daily candles first
daily_plot = fplt.candlestick_ochl(dfd.dropna(), candle_width=5)
daily_plot.colors.update(dict(bull_body='#bfb', bull_shadow='#ada', bear_body='#fbc', bear_shadow='#dab'))
daily_plot.x_offset = 3.1 # resample() gets us start of day, offset +1.1 (gap+off center wick)

# plot high resolution on top
fplt.candlestick_ochl(df[['Open','Close','High','Low']])

# scatter plot correlation between Google and Microsoft stock
df['ret_alphabet']  = df.Close.pct_change()
df['ret_microsoft'] = dfms.Close.pct_change()
dfc = df.dropna().reset_index(drop=True)[['ret_alphabet', 'ret_microsoft']]
fplt.plot(dfc, style='o', color=1, ax=ax2)

# draw least-square line
errfun = lambda arr: [y-arr[0]*x+arr[1] for x,y in zip(dfc.ret_alphabet, dfc.ret_microsoft)]
line = scipy.optimize.least_squares(errfun, [0.01, 0.01]).x
linex = [dfc.ret_alphabet.min(), dfc.ret_alphabet.max()]
liney = [linex[0]*line[0]+line[1], linex[1]*line[0]+line[1]]
fplt.add_line((linex[0],liney[0]), (linex[1],liney[1]), color='#993', ax=ax2)
fplt.add_text((linex[1],liney[1]), 'k=%.2f'%line[0], color='#993', ax=ax2)
fplt.add_legend('Alphabet vs. Microsft 90m correlation', ax=ax2)

fplt.show()

Describe the bug

In my best of opinion, instead of overlay candelles the example code just append the two database.

Expected behavior

I tried something like that but it didn’t work well:

daily_plot = fplt.candlestick_ochl(dfd.dropna(), candle_width=5, ax=ax)
fplt.candlestick_ochl(df[['Open','Close','High','Low']], ax=ax.overlay(scale=1.0, yaxis='linear'))