[Python] Values over bars in barplot not centered

I have the following code, which produces a plot, that plots two dictionaries next to each other, where the y-axis are the days and the y-axis the relative costs. Since the plot is pretty bloated, i want to add the values over each bar, ideally centered. Thus, I added the option ha='center' to my code, but somehow it produces a plot, where the rotated values are not truly centered above the bar. I want the middle of the rotated number centered over the bar. How can i achieve that?

This is my code:

import matplotlib.pyplot as plt
import numpy as np
import random

def plot_relative_undercover_dual(ls1, ls2, demand_dict, days, shifts):
list1, list2 = [], []

for day in range(1, days + 1):
daily_sum1 = sum(ls1.get((day, shift), 0) for shift in range(1, shifts + 1))
daily_sum2 = sum(ls2.get((day, shift), 0) for shift in range(1, shifts + 1))
daily_demand_sum = sum(demand_dict.get((day, shift), 0) for shift in range(1, shifts + 1))

if daily_demand_sum > 0:
relative1 = daily_sum1 / daily_demand_sum
relative2 = daily_sum2 / daily_demand_sum
else:
relative1 = relative2 = 0

list1.append(relative1)
list2.append(relative2)

plt.figure(figsize=(12, 6))

x = np.arange(1, days + 1)
width = 0.35

colors = plt.cm.magma([0.8, 0.2])

bars1 = plt.bar(x - width / 2, list1, width, color=colors[0], alpha=0.7, label='Option 1')
bars2 = plt.bar(x + width / 2, list2, width, color=colors[1], alpha=0.7, label='Option 2')

plt.xlabel('Days', fontsize=14.5, labelpad=15)
plt.ylabel('Relative Costs', fontsize=14.5, labelpad=15)

plt.grid(axis='y', linestyle='--', alpha=0.7)
plt.xticks(x)

def add_value_labels(bars):
for bar in bars:
height = bar.get_height()
plt.text(bar.get_x() + bar.get_width() / 2., height + 0.01,
f'{height:.2%}',
ha='center', va='bottom', rotation=90, fontsize=10, color='black')

add_value_labels(bars1)
add_value_labels(bars2)

# Find the maximum height of all bars
max_height = max(max(bar.get_height() for bar in bars1), max(bar.get_height() for bar in bars2))

# Add a margin above the highest bar for the labels and legend
legend_margin = 0.15 * max_height
plt.ylim(top=max_height + legend_margin)

# Erstelle die Legende mit dünnerem Rahmen
legend = plt.legend(loc='upper left', bbox_to_anchor=(0.02, 0.98), ncol=1, frameon=True, edgecolor='black', facecolor='white', framealpha=1, fontsize=14.5) # Set font size to match axis labels
legend.get_frame().set_linewidth(0.5)

plt.tight_layout()
plt.show()


def generate_random_data(days, shifts, min_value, max_value):
return {(day, shift): round(random.uniform(min_value, max_value), 2) for day in range(1, days + 1) for shift in range(1, shifts + 1)}

demand_dict, u1, u2 = generate_random_data(28, 3, 0, 105), generate_random_data(28, 3, 0, 100), generate_random_data(28, 3, 0, 100)

plot_relative_undercover_dual(u1, u2, demand_dict, 28, 3)


Here you can see the output and the problem more clearly:

Continue reading...