Removing NaN value from legend in matplotlib map plotted from GeoPandas DataFrame

I have a GeoPandas GeoDataFrame which I want to plot against a specific column.

for day in days_list:
    df2 = csv_file[["state_id", "state_name", "tmx_corrl"]].loc[csv_file["days"] == day] 
    new_df = df1.merge(df2, on = 'state_id', how="left")
    new_df.drop(columns=['state_name_y'], inplace = True)
    new_df.rename(columns={"state_name_x": "state_name"}, inplace = True)


    #new_df = pd.merge(df1, df2, on="state_id", how="left")
    #df_new = new_df.dropna(axis="index", how="any")
    df_new = new_df.fillna(0)
    #df_new = new_df.replace(np.nan,0)
    

    cmap = ListedColormap(["red", "orange", "blue", "green"])


    fig, ax = plt.subplots(figsize=(12,8))
    df_new.plot(ax=ax, linewidth=1, color='black')
    

    df_new.plot(
    ax=ax,
    column=new_df["tmx_corrl"],
    linewidth=1,
    cmap=cmap,
    scheme='User_Defined',
    edgecolor='black',
    alpha=0.9,
    classification_kwds={'bins':[0, 0.25, 0.50, 0.75, 1]},
    legend=True
    )
    
    ax.set_title("Day "+ str(day)+" Rainfall")
    
    
    plt.tight_layout()

    plt.savefig("day_{}.png".format(day), dpi=300)

This column have some NaN values, like at column no 7 in this dataframe

which I am removing using several methods

1. Dropping the column with NaN value

    df_new = new_df.dropna(axis="index", how="any")

2. Filling the NaN value to Zero

    df_new = new_df.fillna(0)

3. Replacing the NaN value to Zero

    df_new = new_df.replace(np.nan,0)

NaN value changed to zero

But it is still shows the NaN value in the map legend which I want to remove i.e. the interval " nan – 0.00 " in the legend, as it is altering the whole order of color coding.

Can somebody suggest, what am I missing here?

df_new.plot(
    ax=ax,
    column=df_new["tmx_corrl"], ## change here
    ...
    ...
    )