What I am trying to do:
I am trying to plot some stock data retrieved from the yfinance python package.
The Error:
Matplotlib seems to be creating additional dates on the x-axis. The dates are not present in the Pandas dataframe returned from yfinance, so I’m not exactly sure where they are coming from.
Code to replicate the error:
import yfinance as yf
import matplotlib.pyplot as plt
# change period to "1mo" if running on a friday to view error
data = yf.download("AAPL", period="5d", interval="1d")
plt.bar(data.index, data["Close"])
plt.tick_params("both", labelsize=7)
plt.show()
As you will see matplotlib is adding room for two (empty) additional columns. Which if you print the
dataframe returned from yfinance, are not actual columns with data.
Any help would be greatly appreciated!
I am also curious why this would be the default behavior?
Thanks!
Here is the output of the data variable (type pandas.Dataframe):
Open High Low Close Adj Close Volume
Date
2024-04-12 174.259995 178.360001 174.210007 176.550003 176.550003 101593300
2024-04-15 175.360001 176.630005 172.500000 172.690002 172.690002 73531800
2024-04-16 171.750000 173.759995 168.270004 169.380005 169.380005 73711200
2024-04-17 169.610001 170.649994 168.000000 168.000000 168.000000 50843400
2024-04-18 168.000000 168.639999 167.330002 168.039993 168.039993 9927884
and here is the respective chart:
As you can see, matplotlib is generating room for two columns that, as a result, alter the chart.
I have found a solution to this be using the code:
import yfinance as yf
import matplotlib.pyplot as plt
# change period to "1mo" if running on a friday to view error
data = yf.download("AAPL", period="5d", interval="1d")
dates = data.index.strftime('%Y-%m-%d')
plt.bar(dates, data["Close"])
plt.tick_params("both", labelsize=7)
plt.show()
This replaces creating the bar chart with data.index of type pandas.Index with the dates variable of now type pandas.DatetimeIndex
This creates the following chart:
I still find the behavior of matplotlib to be strange? It at least seems like there should be some type of boolean toggle to turn this feature on or off (maybe there is and I’m just missing it).
