-
Notifications
You must be signed in to change notification settings - Fork 0
/
EDA_Visualization.py
338 lines (285 loc) · 13.5 KB
/
EDA_Visualization.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
import matplotlib.pyplot as plt
import nibabel as nib
import numpy as np
import pandas as pd
import seaborn as sns
import os
import glob
import sklearn
import nilearn as nl
from sklearn.manifold import TSNE
import re
from sklearn.cluster import KMeans
import math
from numpy import dot
from numpy.linalg import norm
import warnings
warnings.filterwarnings('ignore')
"""
This function performs the following task:
-> takes a metadata table containing the mandatory demographics such
as Subject ID, Age, Sex, and the group they belong to.
-> returns the descriptive stats such as no. of subjects in each
group, avg., min, max age, and the gender count
"""
def Comp_Desc_Stats(metadata):
Demographics = pd.DataFrame()
# count subjects in each group
Demographics["Subjects"] = metadata.groupby(["Group"]).count().Subject
Demographics["Avg_Age"] = round(metadata.groupby(["Group"]).mean("Age"),2)
Demographics["Std_Age"] = round(metadata.groupby(["Group"]).std(),2)
Demographics["Min_Age"] = metadata.groupby(["Group"]).min("Age")
Demographics["Max_Age"] = metadata.groupby(["Group"]).max("Age")
# count sex in each group
df = metadata.groupby(["Group","Sex"]).count().Subject.to_frame(name = 'Sex_Count').reset_index()
Demographics["Male"] = df.loc[df["Sex"]=="M", "Sex_Count"].values
Demographics["Female"] = df.loc[df["Sex"]=="F", "Sex_Count"].values
Demographics["Group"] = Demographics.index
Demographics.reset_index(inplace=True, drop=True)
total_sub = metadata.loc[:,"Subject"].count()
avg_age = round(metadata.loc[:,"Age"].mean(),2)
std_age = round(metadata.loc[:,"Age"].std(),2)
min_age = metadata.loc[:,"Age"].min()
max_age = metadata.loc[:,"Age"].max()
male_count = metadata.groupby(["Sex"]).count().Subject.values[1]
female_count = metadata.groupby(["Sex"]).count().Subject.values[0]
overall_ser = pd.Series([ total_sub, avg_age, std_age, min_age, max_age, male_count, female_count, "Overall"])
overall = pd.DataFrame(overall_ser).transpose()
overall.columns = Demographics.columns.values
Demographics = pd.concat([Demographics, overall], ignore_index=True)
return Demographics
"""
The below function does the following task:
---------------------------------------------
Creates an histogram for a continuous demographic variable such as age
"""
def Plot_Age_Hist(Ages, title, x_loc, path):
mu = round(Ages.loc[:,"Age"].mean(),2)
std = round(Ages.loc[:,"Age"].std(),2)
min_age = round(Ages.loc[:,"Age"].min(), 0)
max_age = round(Ages.loc[:,"Age"].max(),0)
Ages["Age_Bins"] = pd.cut(x=Ages["Age"], bins=10)
Ages["Age_Bins"] = Ages["Age_Bins"].apply(lambda x: pd.Interval(left=int(round(x.left)), right=int(round(x.right))))
Ages_Piv = Ages.pivot_table("Age", index="Age_Bins", aggfunc='count', fill_value=0)
Ages_Piv.plot(kind="bar", stacked=True, width=0.9, color="cadetblue", edgecolor="black", linewidth=0.5)
# number of fixed locators
x = np.arange(0,10)
label = np.arange(min_age,max_age+1,step=(max_age-min_age)/10 )
labels = []
for i in np.arange(0,len(label)-1):
# get the middle point of the age interval
labels.append(round((label[i]+label[i+1])/2,None))
print(labels, x)
plt.xticks(ticks=x, labels=labels, rotation=None ) # Set label locations.
# if axis_labels==True:
plt.xlabel("Ages")
plt.ylabel("No. of participants")
plt.tight_layout()
plt.title(title)
plt.tight_layout()
plt.text(x_loc, Ages_Piv.max().max()-5, r'$\mu='+str(mu)+',\ \sigma='+str(std)+'$')
plt.savefig(path + title+ "_age_hist_with_title.png", bbox_inches='tight', dpi=300)
"==============================================================================="
"""
The below function does the following task:
---------------------input---------------------
1. A dataframe with two columns (age and gender)
2. title of the histogram to appear on the graph
3. Path to save the figure
---------------------output--------------------
An stacked bar chart showing the age and gender frequency in each interval
"""
def Plot_Age_Gend_Hist(sub_demographics, title, x_loc, y_loc, path):
mu = round(sub_demographics.loc[:,"Age"].mean(),1)
sigma = round(sub_demographics.loc[:,"Age"].std(),1)
# create an age bin column for each row where an age belongs to
sub_demographics["Age_bins"] = pd.cut(x=sub_demographics.loc[:,"Age"], bins=10)
sub_demographics["Age_bins"] = sub_demographics.loc[:,"Age_bins"].apply(lambda x: pd.Interval(left=int(round(x.left)), right=int(round(x.right))))
# get the count of male and female subjects in all age intervals
piv_age_sex = sub_demographics.pivot_table("Age", index="Age_bins", columns="Gender", aggfunc='count', fill_value=0)
min_age = round(sub_demographics.loc[:,"Age"].min(), 0)
max_age = round(sub_demographics.loc[:,"Age"].max(),0)
# number of points or locations on the x axis
x = np.arange(0,10)
label = np.arange(min_age,max_age+1,step=(max_age-min_age)/10 )
labels = []
for i in np.arange(0,len(label)-1):
# get the middle point of the age interval
labels.append(round((label[i]+label[i+1])/2,None))
print(labels)
# set the column names (to appear as legend)
piv_age_sex.columns = ["Male", "Female"]
ax = piv_age_sex.plot(kind="bar", stacked=True, color=["azure", "white"], fill=True, width=0.8, edgecolor="gray", hatch="/.")
# get all bars in the plot
bars = ax.patches
patterns = ['', '.'] # set hatch patterns in the correct order
hatches = [] # list for hatches in the order of the bars
for h in patterns: # loop over patterns to create bar-ordered hatches
for i in range(int(len(bars) / len(patterns))):
hatches.append(h)
for bar, hatch in zip(bars, hatches): # loop over bars and hatches to set hatches in correct order
bar.set_hatch(hatch)
# generate legend. this is important to set explicitly, otherwise no hatches will be shown!
ax.legend()
plt.rcParams.update({'font.size': 14})
plt.xticks(ticks=x, labels=labels, rotation=None)
plt.title(title)
plt.xlabel("Ages ", fontsize=14)
plt.ylabel("No. of participants", fontsize=14)
# plt.grid(linewidth=0.5)
plt.tight_layout()
plt.text(x_loc, y_loc, r'$\mu='+str(mu)+',\ \sigma='+str(sigma)+'$')
plt.savefig(path + title+ "_age_gender_hist_with_title.png", bbox_inches='tight', dpi=300)
"=================================================================================="
"""
The below function does the following task:
-----------------------------------------
Visualize the relationship between chronological ages and the estimated
brain ages using scatter plots
-----------------------------------------
Takes the actual and predicted brain ages as input
"""
" ========= when both healthy and cases ============"
def Plot_true_vs_pred_grouped(Y_pred_hc,Y_test_hc, Y_pred_case, Y_test_case, title, R_sqq, path):
plt.figure(figsize=(8, 6))
plt.rcParams.update({'font.size': 20})
sp_names = [ "Healthy Control", "Brain Tumor"]
sct1 = plt.scatter(Y_test_hc, Y_pred_hc, c='tan', s=50, marker="v")
sct2 = plt.scatter(Y_test_case, Y_pred_case, c='cadetblue', s=50, marker="x")
plt.legend((sct1, sct2), labels=sp_names)
Plot_Scatter(R_sqq,title, path)
"========== healthy controls only ================"
def Plot_true_vs_pred(Y_test_hc, Y_pred_hc, title, R_sqq, path):
plt.figure(figsize=(8, 6))
sct1 = plt.scatter(Y_test_hc, Y_pred_hc, c='saddlebrown', s=20)
# plt.legend((sct1, sct2), labels=sp_names)
Plot_Scatter(R_sqq, title, path)
"========= Reusable code block ==================="
def Plot_Scatter(R_sqq,title,path):
plt.yscale('linear')
plt.xscale('linear')
plt.rcParams.update({'font.size': 16})
p1 = 5
p2 = 80
plt.plot([p1, p2], [p1, p2], 'dimgray')
plt.plot([p1+10, p2+10], [p1, p2], 'dimgray', linestyle="--")
plt.plot([p1, p2], [p1+10, p2+10], 'dimgray', linestyle="--")
plt.text(10, 60, "$R^2 = " + str(R_sqq) +"$", fontsize =16)
plt.xlabel("Chronological Age")
plt.ylabel("Estimated Brain Age")
plt.title(title)
plt.tight_layout()
plt.savefig(path+"true_vs_pred_scat_plot.png", dpi=300)
def Plot_true_vs_BAG_grouped(Y_pred_hc,Y_test_hc, Y_pred_case, Y_test_case, title, path):
plt.figure(figsize=(8, 6))
plt.rcParams.update({'font.size': 18})
fig, ax = plt.subplots()
BAG_test = (Y_pred_hc-np.array(Y_test_hc).reshape(1,-1)).transpose()
BAG_bt = (Y_pred_case-np.array(Y_test_case).reshape(1,-1)).transpose()
sp_names = [ "Healthy Control", "Brain Tumor"]
sct1 = plt.scatter( Y_test_hc, BAG_test, marker="v", c="tan", s=50)
sct2 = plt.scatter( Y_test_case, BAG_bt, marker="x", c="cadetblue", s=50)
plt.legend((sct1, sct2), labels=sp_names)
Y_test_arr = np.array(Y_test_hc).flatten()
BAG_test_arr = np.array(BAG_test).flatten()
# calc the trendline
z = np.polyfit(Y_test_arr, BAG_test_arr, 1)
p = np.poly1d(z)
plt.plot(Y_test_arr,p(Y_test_arr),"r--", color="tan")
Y_bt_arr = np.array(Y_test_case).flatten()
BAG_bt_arr = np.array(BAG_bt).flatten()
# calc the trendline
z = np.polyfit(Y_bt_arr, BAG_bt_arr, 1)
p = np.poly1d(z)
plt.plot(Y_bt_arr,p(Y_bt_arr),"r--", color="cadetblue")
plt.ylabel("BAG (years)")
plt.xlabel("Chronological Age")
plt.axhline(0, color='gray', linestyle="dotted")
plt.tight_layout()
plt.savefig(path+"true_vs_BAG_scat_plot.png", dpi=300)
####### Plot the age-gap box plots for different features
def Box_plot_age_gap_diff_feat(Y_test, Pred_brain_age, xlabels, path):
plt.rcParams.update({'font.size': 15})
fig, ax = plt.subplots()
bp = plt.boxplot(Pred_brain_age-Y_test.reshape(-1,1), 1, "", positions=np.arange(1,15,2), widths=1,
patch_artist=True, boxprops=dict(facecolor="white"))
for box in bp['boxes']:
# change outline color
box.set(linewidth=1)
# change hatch
box.set(hatch = '\\')
plt.ylabel("Brain-EAD (years)")
if (xlabels==True):
ax.set_xticklabels(["CAT12 ROI", "Desikan ROI", "Destrieux ROI", "CAT12 ROI + Desikan ROI",
"CAT12 ROI + Destrieux ROI","Desikan ROI + Destrieux ROI", "All region wise" ], rotation=90)
else:
plt.xticks([1], [''])
ax.set_xlim(0,15)
plt.axhline(0, color='gray', linestyle="dotted")
plt.savefig(path + "_no_labels.png", bbox_inches='tight', dpi=300)
####### Plot the age-gap box plots for different features
def Box_plot_ages(Y_test, Pred_brain_age, xlabels, path):
plt.rcParams.update({'font.size': 15})
fig, ax = plt.subplots()
bp = plt.boxplot(Pred_brain_age-Y_test.reshape(-1,1), 1, "", positions=np.arange(1,15,2), widths=1,
patch_artist=True, boxprops=dict(facecolor="white"))
for box in bp['boxes']:
# change outline color
box.set(linewidth=1)
# change hatch
box.set(hatch = '\\')
plt.ylabel("Brain-EAD (years)")
if (xlabels==True):
ax.set_xticklabels(["CAT12 ROI", "Desikan ROI", "Destrieux ROI", "CAT12 ROI + Desikan ROI",
"CAT12 ROI + Destrieux ROI","Desikan ROI + Destrieux ROI", "All region wise" ], rotation=90)
else:
plt.xticks([1], [''])
ax.set_xlim(0,15)
plt.axhline(0, color='gray', linestyle="dotted")
plt.savefig(path + "_no_labels.png", bbox_inches='tight', dpi=300)
def Box_Plot_Age_Gap(Y_pred_test, Y_test, Y_pred_bt, Y_bt, path):
plt.rcParams.update({'font.size': 15})
fig, ax = plt.subplots()
BAG_test = (Y_pred_test-np.array(Y_test).reshape(1,-1)).transpose()
BAG_bt = (Y_pred_bt-np.array(Y_bt).reshape(1,-1)).transpose()
# combine the two arrays with unequal lengths
data = np.array([BAG_test, BAG_bt])
# box plot of two arrays or columns with unequal lengths
bp = plt.boxplot(data, patch_artist=True, widths=0.5)
i=0
for box in bp['boxes']: # loop through the boxes
# apply different conditions to the boxes
if (i==0):
# change outline color
box.set(facecolor="wheat")
else:
box.set(facecolor="cadetblue")
i = i+1
plt.ylabel("BAG (years)")
ax.set_xticklabels(["Healthy Control", "Brain Tumor"])
plt.axhline(0, color='gray', linestyle="dotted")
plt.savefig(path + "_age_gap_box_plot.png", bbox_inches='tight', dpi=300)
def Plot_Survival_vs_BAG(Survival_days, BAG_bt, path):
plt.rcParams.update({'font.size': 15})
fig, ax = plt.subplots()
BAG_bt = abs(BAG_bt)
# create a dataframe on with BAG and Survival_days as columns and sort it
x = pd.DataFrame({"BAG": pd.Series(BAG_bt), "Survival_days": Survival_days.astype(int)})
x = x.sort_values(by=["Survival_days"], ascending=False)
plt.scatter(x.loc[:,"BAG"], x.loc[:,"Survival_days"], marker="x", c="cadetblue")
# insert comma when the legend label text is truncated or sliced showing only first alphabet
plt.legend(("Brain Tumor",))
BAG_bt_arr = np.array(x.loc[:,"BAG"], dtype=float).reshape(60)
Survival_arr = np.array(x.loc[:,"Survival_days"], dtype=float).reshape(60)
# calc the trendline
z = np.polyfit(BAG_bt_arr, Survival_arr, 1)
p = np.poly1d(z)
plt.plot(BAG_bt_arr, p(BAG_bt_arr),"r--", color="tan")
# set y tick labels after a certain 'n' here n=1
# used when length of the points is very much
ax.set_yticks(ax.get_yticks()[::1])
plt.ylim(0,2000)
plt.xlabel("BAG (years)")
plt.ylabel("Survival (days)")
plt.tight_layout()
plt.savefig(path+"BAG_vs_survival_scat_plot.png", dpi=300, bbox_inches='tight')