Code output demonstration
Multiple color schemes
Code explanation
Part One
Library imports and font settings
# =========================================================================================# ====================================== 1. Environment Setup =======================================# =========================================================================================import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from matplotlib.patches import Polygon
from scipy.stats import gaussian_kde, ttest_1samp
import matplotlib.patches as mpatches
import os
plt.rcParams['font.family'] = 'serif'
plt.rcParams['font.serif'] = ['Times New Roman']
plt.rcParams['axes.unicode_minus'] = False
Part Two
Color library setup and color scheme selection
# =========================================================================================# ====================================== 2. Color Library Setup ======================================# =========================================================================================COLOR_LIBRARY = { 1: {'T1': '#005A8D', 'T2': '#CFA900', 'T3': '#006400', 'T4': '#8B0000'},}
SELECTED_SCHEME=2
colors_map = COLOR_LIBRARY[SELECTED_SCHEME]
Part Three
Analysis function to calculate statistical significance and set significance markers based on analysis results
# =========================================================================================# ====================================== 3. Analysis Function ======================================# =========================================================================================
def get_significance_label(period_values, baseline_value):
period_values = np.array(period_values) # Convert input list to numpy array
period_values = period_values[~np.isnan(period_values)] # Remove NaN values
if len(period_values) < 2: # If valid data points are less than 2
return "ns" # Return no significance marker
t_stat, p_val = ttest_1samp(period_values, baseline_value) # Perform one-sample T-test
# Set significance marker based on analysis results
if p_val < 0.0001:
sig = "****"
elif p_val < 0.001:
sig = "***"
elif p_val < 0.01:
sig = "**"
elif p_val < 0.05:
sig = "*"
else:
sig = "ns"
# Set significance value label based on analysis results
if p_val < 0.0001:
return f"p < 0.0001{sig}"
else:
return f"p = {p_val:.4f}{sig}"
Part Four
Function to draw the raincloud plot for individual subplots in the main figure
# =========================================================================================# ====================================== 5. Individual Raincloud Plot Drawing Function ======================================# =========================================================================================
def draw_raincloud_radial(ax, angle, values, color, width_angle):
cloud_offset = width_angle * 0.15 # Calculate angle offset for cloud density plot
cloud_width_max = width_angle * 0.3 # Maximum angle width for cloud density plot
scatter_jitter_width = width_angle * 0.04 # Random jitter width for scatter plot
box_offset = -width_angle * 0.15 # Angle offset for box plot
box_width = width_angle * 0.22 # Define width of box plot
kde = gaussian_kde(values) # Perform Gaussian kernel density estimation on data
r_grid = np.linspace(0, 150, 200) # Create grid points in the radial direction
kde_vals = kde(r_grid) # Calculate density values at grid points
max_kde = kde_vals.max() # Get maximum density value
scaling_factor = cloud_width_max / max_kde if max_kde > 0 else 0 # Calculate scaling factor to fit maximum width
theta_base = angle + cloud_offset # Determine base angle for cloud density plot
theta_curve = theta_base + kde_vals * scaling_factor # Calculate angle curve for cloud density plot edges based on density values
verts = list(zip(theta_curve, r_grid)) + list(zip([theta_base] * len(r_grid), r_grid[::-1])) # Construct polygon vertex coordinates
# Create cloud density plot
poly = Polygon(verts,
facecolor=color,
edgecolor='none',
alpha=1,
zorder=1)
ax.add_patch(poly) # Add cloud density plot to axes
# Box plot
q1 = np.percentile(values, 25) # Lower quartile
med = np.median(values) # Median
q3 = np.percentile(values, 75) # Upper quartile
iqr = q3 - q1 # Interquartile range
whisk_min = np.min(values[values >= q1 - 1.5 * iqr]) # Lower whisker
whisk_max = np.max(values[values <= q3 + 1.5 * iqr]) # Upper whisker
box_center_angle = angle + box_offset # Center angle for box plot
# Draw box
ax.bar(x=box_center_angle,
height=q3 - q1,
bottom=q1,
width=box_width,
color=color,
edgecolor='black',
linewidth=1.5,
alpha=1,
zorder=10)
# Draw median line
ax.plot([box_center_angle - box_width / 2,
box_center_angle + box_width / 2],
[med, med],
color='black',
lw=1.5,
zorder=11)
# Draw lower whisker line
ax.plot([box_center_angle,
box_center_angle],
[whisk_min, q1],
color='black',
lw=1.5,
zorder=9)
# Draw upper whisker line
ax.plot([box_center_angle,
box_center_angle],
[q3, whisk_max],
color='black',
lw=1.5,
zorder=9)
Part Five
Function to draw significance arcs
# =========================================================================================# ====================================== 6. Significance Arc Drawing Function ======================================# =========================================================================================
def add_significance_arc_dynamic(ax, start_angle, end_angle, text, color):
r_arc = 158 # Radius height for significance arc
theta = np.linspace(start_angle, end_angle, 100) # Angle sequence for arc
# Draw arc
ax.plot(theta,
[r_arc] * len(theta),
color=color,
lw=2)
ax.plot([start_angle, start_angle], [r_arc - 5, r_arc], color=color, lw=2) # Starting vertical short line
ax.plot([end_angle, end_angle], [r_arc - 5, r_arc], color=color, lw=2) # Ending vertical short line
mid_angle = (start_angle + end_angle) / 2 # Midpoint angle for arc
deg = np.degrees(mid_angle) # Convert radians to degrees
text_rot = 270 - deg + 90 # Text rotation angle
# Add significance text label
ax.text(mid_angle,
r_arc + 12,
text,
rotation=text_rot,
ha='center',
va='center',
fontsize=12,
fontweight='bold',
color='black')
Part Six
Function to draw the radial raincloud plot/main figure
# =========================================================================================# ====================================== 7. Main Plotting Function ======================================# =========================================================================================
def plot_radial_chart(data_dict, groups_list, periods_list, global_mean, save_dir):
fig, ax = plt.subplots(figsize=(12, 12), subplot_kw={'projection': 'polar'}) # Create canvas
ax.set_theta_zero_location("N") # Set zero degree position for polar coordinates to North
ax.set_theta_direction(-1) # Set angle increase direction to clockwise
ax.set_ylim(-150, 180) # Set radial range
num_groups = len(groups_list) # Get total number of groups
gap_degrees = 40 # Set gap angle size
gap_radians = np.radians(gap_degrees) # Convert gap angle to radians
start_angle = gap_radians / 2 # Starting angle
end_angle = 2 * np.pi - (gap_radians / 2) # Ending angle
angles = np.linspace(start_angle, end_angle, num_groups) # Center angles for each group
width_per_group = (end_angle - start_angle) / (num_groups - 1) * 0.55 # Angle width allocated for each group
# Loop to draw each group
for i, (g_name, angle) in enumerate(zip(groups_list, angles)): # Iterate through each group and corresponding angle
g_data = data_dict[g_name] # Get data for that group
vals = g_data['vals'] # Get value array
period = g_data['period'] # Get period label
# Get color based on color scheme
color = colors_map.get(period, '#555555')
# Call function to draw raincloud plot
draw_raincloud_radial(ax, angle, vals, color, width_angle=width_per_group)
label_r = -20 # Radius position for group name label
rot_angle = 90 - np.degrees(angle) # Rotation angle for label
# Add group name
ax.text(angle,
label_r,
g_name,
rotation=rot_angle,
ha='right',
va='center',
fontsize=13,
fontweight='bold',
color='black',
rotation_mode='anchor')
ax.grid(False) # Remove default grid
ax.spines['polar'].set_visible(False) # Remove polar spine
ax.set_xticks([]) # Remove angle ticks
ax.set_yticks([]) # Remove radial ticks
# To prevent raincloud plot from exceeding the circle
gap_degrees = 30 # Set gap angle size
gap_radians = np.radians(gap_degrees) # Convert gap angle to radians
start_angle = gap_radians / 2 # Starting angle
end_angle = 2 * np.pi - (gap_radians / 2) # Ending angle
grid_angles = np.linspace(start_angle, end_angle, 200) # Generate angle sequence for circle
for y in [0, 60, 120, 180]: # Iterate through radial positions to draw circles
line_style = '-' if y == 0 else '--' # Solid line for 0 position, dashed for others
if y == 180:
linewidth = 4.0 # Line for outermost circle
elif y == 0:
linewidth = 2.0 # Line for 0
else:
linewidth = 2.0 # Middle line
# Get unique period list and sort by numeric suffix
unique_periods = sorted(list(set(periods_list)), key=lambda x: int(x[1:]) if (isinstance(x, str) and x[1:].isdigit()) else x)
period_indices = {p: [] for p in unique_periods} # Initialize dictionary to store indices for each period
for idx, p in enumerate(periods_list): # Iterate through all period lists
period_indices[p].append(idx) # Record index for each period
for p in unique_periods: # Iterate through each unique period
idxs = period_indices[p] # Get all indices for that period
start_idx = idxs[0] # Get starting index
end_idx = idxs[-1] # Get ending index
current_period_vals = [] # Initialize list for all values for current period
for i in idxs: # Iterate through each index within that period
g_name = groups_list[i] # Get group name
current_period_vals.extend(data_dict[g_name]['vals']) # Add that group's values to the list
sig_label = get_significance_label(current_period_vals, global_mean) # Calculate and get significance label
c = colors_map.get(p, 'black') # Get color for that period
add_significance_arc_dynamic(ax, angles[start_idx], angles[end_idx], sig_label, c) # Draw significance arc and label
legend_handles = [mpatches.Patch(color=colors_map.get(p, 'gray'), label=p) for p in unique_periods] # Create legend handle list
# Add legend
leg = ax.legend(handles=legend_handles,
loc='center',
title="Period",
frameon=False,
fontsize=18,
bbox_to_anchor=(0.5, 0.5),
bbox_transform=ax.transAxes)
leg.get_title().set_fontsize(20) # Set legend title
leg.get_title().set_fontweight('bold') # Set legend title
plt.setp(leg.get_texts(), fontweight='bold') # Set legend text
plt.tight_layout() # Auto-adjust
Part Seven
Execution section, including data reading, analysis, plotting, and saving
# =========================================================================================# ====================================== 8. Execution Section ======================================# =========================================================================================
if __name__ == "__main__":
excel_file = r'Data.xlsx' # Original data
output_folder = r'1125-Radial Raincloud Plot' # Output path
df_raw = pd.read_excel(excel_file) # Read data
required_cols = ['Group', 'Period', 'Value'] # Required column names
groups_list = df_raw['Group'].unique().tolist() # Get unique group name list
all_values_raw = df_raw['Value'].values # Get all value data
global_mean = np.mean(all_values_raw) # Calculate global mean as baseline
print(f"Global Mean (Baseline): {global_mean:.2f}") # Print global mean
# Build data dictionary
data_dict = {} # Initialize data dictionary
periods_list = [] # Initialize period order list
for g in groups_list: # Iterate through each group name
sub_df = df_raw[df_raw['Group'] == g] # Filter current group's data
p_val = sub_df['Period'].iloc[0] # Get current group's corresponding period identifier
vals = sub_df['Value'].values # Get current group's value array
data_dict[g] = { 'vals': vals, # Store values 'period': p_val # Store period }
periods_list.append(p_val) # Add period to list
plot_radial_chart( data_dict=data_dict, # Data groups_list=groups_list, # Group names periods_list=periods_list, # Periods global_mean=global_mean, # Global mean save_dir=output_folder # Save path )
How to apply?
1. Choose the color scheme you want to use:
SELECTED_SCHEME=22. Set the input path for the original data:
excel_file = r'Data.xlsx' # Original data3. Set the save address for the plotting results:
output_folder = r'1125-Radial Raincloud Plot'4. Define the column names for the data you want to use:
required_cols = ['Group', 'Period', 'Value']
Recommendations
Journal image reproduction | Python drawing of 2D partial dependence PDP plotsJournal reproduction | Python drawing of single feature dependence plots based on SHAP analysis and GAM model fittingJournal image reproduction | Python drawing of gradient color SHAP feature importance combination plots (bar chart + hexagon plot)Journal reproduction | Python drawing of SHAP feature importance overview plots, dependence plots, and dual feature interaction effect SHAP plots, unlocking the ultimate secrets of XGBoost modelsJournal image reproduction | Python drawing of SHAP importance hexagon plots + single feature dependence plots + interaction effect intensity bubble plots + interaction effect dependence plots (regression + binary classification + classification)
Acquisition method
The purchase channel for the paid collection of codes and data has been opened, and updates will continue. If needed, pleasemessage me privately for detailed information. Note that only practice data and code files will be shared, and no Q&A service will be provided. The code files already contain complete comments for each line of code. The codes in the paid collection can still be obtained by recommending + liking ten articles, but the number of times is limited (up to 10 articles). Previous free articles will still be free, and there will also be free articles in the future, so please ensure you really need it before purchasing!!!