Update basketball_app.py

[odagayev]

Changing the data type to str in order to avoid the "("Expected bytes, got a 'int' object", 'Conversion failed for column FG% with type object')" error by the st.dataframe on line 45.

This is a new error introduced in streamlit v.85.

[jdhauswirth]

Proposed (2) Minor .py File Changes

https://github.com/jdhauswirth/streamlit_freecodecamp/blob/main/app_3_eda_basketball/basketball_app.py

Hello DataProf,

I sincerely appreciate and thank you for great content and tutorials. As noted in the YT video (https://youtu.be/JwSS70SZdyM) comments section, this script has presented an error for several users. I hope you will please consider the following 2 small revisions to resolve 1 error and 1 warning with App03 - basketball_app.py.

1). To resolve error "("Expected bytes, got a 'int' object", 'Conversion failed for column FG% with type object')" - Please consider a small change to wrap/typecast df_selected_team in line 40 as string prior to use of df_selected_team in line 44. (NOTE* - The existing PR (thank you to odagayev) that adds another line of code however I wish to suggest wrapping the existing assignment to avoid addition of another line of code.)

Proposed Change of Line 40
from
df_selected_team = playerstats[(playerstats.Tm.isin(selected_team)) & (playerstats.Pos.isin(selected_pos))]
to
df_selected_team = (playerstats[(playerstats.Tm.isin(selected_team)) & (playerstats.Pos.isin(selected_pos))]).astype(str)

2). To resolve deprecation warning when clicking the "Intercorrelation Heatmap" button - PyplotGlobalUseWarning: You are calling st.pyplot() without any arguments. Please consider 1 character modification to line 68 to explicitly pass f (figure) as a parameter to resolve warning.

Proposed Change of Line 68
from
st.pyplot()
to
st.pyplot(f)

Surely if these are erroneous, present issue(s) or there is a simpler solution, please. I thank you in advance for your time and consideration.

Sincerely,

JDHaus

[Luismbpr]

Hello.
I propose the following minor changes. This is to make sure there are no conflicts regarding strings on the correlation so the heat map is displayed.

1. Dropping the non-numerical columns (Player, Position and Team) from the re-uploaded DataFrame.
2. Adding the st.pyplot(f) so the figure shows correctly.

The code would be:

# Heatmap
if st.button('Intercorrelation Heatmap'):
    st.header('Intercorrelation Matrix Heatmap')
    df_selected_team.to_csv('output.csv',index=False)
    df = pd.read_csv('output.csv')
    #Dropping Non-Numerical (string) Columns
    df_selected_htmp = df.drop(columns=['Player', 'Pos', 'Tm'], axis=1)
#    corr = df.corr()
    corr = df_selected_htmp.corr()
    mask = np.zeros_like(corr)
    mask[np.triu_indices_from(mask)] = True
    with sns.axes_style("white"):
        f, ax = plt.subplots(figsize=(7, 5))
        ax = sns.heatmap(corr, mask=mask, vmax=1, square=True)
    st.pyplot(f)