In python matplotlib, when plotting a multi-bar graph can I skip bars if the value is NaN so they do not appear?

Question

I'm coding in Python using the library matplotlib to make a multi-bar graph, like a double-bar graph but I have 7 bars per category. I'm not sure what I have is necessarily the best way for me to plot this data, but it's working.

However some of the data for those 7 bars per group are missing, where the value is 0 or NaN. This is creating gaps in the data where the bar is missing as well.

Is it possible to modify the graphing code to skip the bars that have a value of 0 without messing with the indents and spacing, etc?

Here is the data I'm using. I've posted it here as a dictionary to copy over but my format is using a DataFrame.

# Libraries
import pandas as pd 
import matplotlib.pyplot as plt 
import numpy as np 

# Dictionary:
bardict = {"B": [0.14,0.12,0.02,0.02,nan,nan,nan],
           "L": [0.08,0.14,0.06,0.06,0.07,0.12,0.08],
           "M": [0.08,0.09,0.07,0.08,0.22,0.15,0.06],
           "C": [0.11,0.10,0.13,0.35,nan,0.11,0.21],
           "S": [nan,0.11,0.17,0.46,0.09,0.10,0.08],
           "W": [0.12,0.09,0.29,0.63,0.10,0.38,0.26]}

# DataFrame saved under name bar_C:
      B     C     L     M     S     W
0  0.14  0.11  0.08  0.08   NaN  0.12
1  0.12  0.10  0.14  0.09  0.11  0.09
2  0.02  0.13  0.06  0.07  0.17  0.29
3  0.02  0.35  0.06  0.08  0.46  0.63
4   NaN   NaN  0.07  0.22  0.09  0.10
5   NaN  0.11  0.12  0.15  0.10  0.38
6   NaN  0.21  0.08  0.06  0.08  0.26

And finally, here is my code so far for making the bar graph: (Note* - I don't want to arrange the bar graph by value, I want them kept in the same order listed (B, C, L, M...)

# Convert dictionary to DataFrame
bar_C = pd.DataFrame(bardict)


fig, ax = plt.subplots(figsize=(10, 6))

N = 6
ind = np.arange(N) #*
width = 0.55

num3_vals = bar_C.loc[0]
num3 = plt.bar(ind*N+(width*0), num3_vals, width, color = 'mediumturquoise')

num4_vals = bar_C.loc[1]
num4 = plt.bar(ind*N+(width*1), num4_vals, width, color='darkred')

num5_vals = bar_C.loc[2]
num5 = plt.bar(ind*N+(width*2), num5_vals, width, color='lightgreen')

num6_vals = bar_C.loc[3]
num6 = plt.bar(ind*N+(width*3), num6_vals, width, color='purple')

num7_vals = bar_C.loc[4]
num7 = plt.bar(ind*N+(width*4), num7_vals, width, color='gray')
print(num7_vals)

num8_vals = bar_C.loc[5]
num8 = plt.bar(ind*N+(width*5), num8_vals, width, color='orange')

num9_vals = bar_C.loc[6]
num9 = plt.bar(ind*N+(width*6), num9_vals, width, color='darkblue')

plt.xlabel("Group")
plt.ylabel('Value')
plt.title("Multi-Bar Graph")

ax = plt.gca()

plt.xticks(ticks = (ind*N),labels = ['B','C','L','M','S','W'])
plt.legend((num3, num4, num5, num6, num7, num8, num9), 
           ('3','4','5','6','7','8','9'),
           loc='upper left')
plt.show()

Output:

B     NaN
L    0.07
M    0.22
C     NaN
S    0.09
W    0.10
Name: 4, dtype: float64

Pic:

I'm a brand new user - so firstly thanks in advance and secondly, please let me know if there's any formatting errors.

I am hoping for all the bars with data, AKA not being zeroes, to be plotted as available. It worked except it left a blank spot where a bar would be for the zero values. I'd like to clean it up so that the empty spaces are gone without having to delete the data.

I tried using fillna() to make everything 0 instead of NaN, but this didn't change the graph in any way.

In python matplotlib, when plotting a multi-bar graph can I skip bars if the value is NaN so they do not appear?

Answers (1)

Related Questions