-
Notifications
You must be signed in to change notification settings - Fork 1
/
grouped_ops_demo.qmd
255 lines (200 loc) · 8.4 KB
/
grouped_ops_demo.qmd
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
# Grouping and Summarizing Data in Python with pandas
```{python}
import pandas as pd
import numpy as np
###############################################################################
# Pandas GroupBy Operations Demo
###############################################################################
###############################################################################
# 1. Data Creation
###############################################################################
# Create sample data
df = pd.DataFrame(
{
"id": range(1, 21),
"city": np.random.choice(["NY", "LA", "SF"], 20),
"age_group": np.random.choice(["child", "adult"], 20),
"purchases": np.random.randint(20, 100, 20),
"returns": np.random.randint(1, 10, 20),
}
)
###############################################################################
# 2. Simple Grouped Aggregation
###############################################################################
# Using agg(): Get mean purchase count per age group
df.groupby("age_group").agg(purchases_mean=("purchases", "mean"))
# Multiple groups
df.groupby(["age_group", "city"]).agg(purchases_mean=("purchases", "mean"))
# Note: The dictionary method {column: aggregation} is not preferred due to lack of
# control over column names and potential issues with nested dataframes.
###############################################################################
# 3. Complex Grouped Aggregation
###############################################################################
(
df.groupby("age_group")
.agg(
purchases_mean=("purchases", "mean"),
purchases_max=("purchases", "max"),
sd_population=("purchases", "std"),
sd_sample=("purchases", lambda x: x.std(ddof=1)),
purchase_sum=("purchases", "sum"),
return_sum=("returns", "sum"),
)
.assign(purchase_return_ratio=lambda x: x.purchase_sum / x.return_sum)
)
###############################################################################
# 4. Grouped Assignment
###############################################################################
# Simple: Add column with difference from group mean using assign
(
df.assign(
mean_purchase=df.groupby("age_group")["purchases"].transform("mean"),
diff_mean=lambda x: (x.purchases - x.mean_purchase).abs(),
categ=lambda x: np.where(x.diff_mean > 30, "extreme", "normal"),
)
)
###############################################################################
# 5. Grouped Arrange (Sorting)
###############################################################################
# Sort by age_group, then by purchases descending
df.sort_values(["age_group", "purchases"], ascending=[True, False])
###############################################################################
# 6. Grouped Filtering
###############################################################################
# Get rows with minimum purchases for each age group using assign and query
(
# first create a column with the min purchase for each age group
df.assign(min_purchase=df.groupby("age_group")["purchases"].transform("min"))
# then query
.query("purchases == min_purchase")
)
# 6.3 Compound: Get rows with both minimum and maximum purchases for each age group
# Using assign and query
(
df.assign(
min_purchase=df.groupby("age_group")["purchases"].transform("min"),
max_purchase=df.groupby("age_group")["purchases"].transform("max"),
)
.query("purchases == min_purchase or purchases == max_purchase")
.drop(columns=["min_purchase", "max_purchase"])
)
```
```{python}
## Recreate this whole pipeline in Python pandas using modern, easily readable concise code.
library(tidyverse)
library(lubridate)
# Generate 200 random records
lab_dat <- tibble(
hospital_num = sample(paste0("P", 1:50), 200, replace = TRUE),
date_visit = seq(as.Date("2020-01-01"), as.Date("2024-08-16"), by = "day")[sample(1:1689, 200, replace = TRUE)]) %>%
arrange(hospital_num, date_visit)
viral_load_tests_plot <-
lab_dat %>%
## sample groups
filter(hospital_num %in% sample(unique(hospital_num), 20)) %>%
## months since last test
group_by(hospital_num) %>%
arrange(date_visit, .by_group = T) %>%
mutate(days_since_last_test = date_visit - lag(date_visit, n = 1, default = NA)) %>%
ungroup() %>%
select(hospital_num, days_since_last_test, date_visit) %>%
## longest month since last test (for arranging ggplot)
group_by(hospital_num) %>%
mutate(longest_test_interval = max(days_since_last_test, na.rm = T)) %>%
ungroup() %>%
## factor order
mutate(hospital_num = fct_reorder(hospital_num, longest_test_interval)) %>%
## plot
ggplot(aes(group = hospital_num, x = date_visit, y = hospital_num)) +
geom_point() +
geom_line() +
theme(legend.position = "none")
viral_load_tests_plot
```
```{python}
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from datetime import datetime, timedelta
# Generate 200 random records
start_date = datetime(2020, 1, 1)
end_date = datetime(2024, 8, 16)
date_range = pd.date_range(start=start_date, end=end_date)
lab_dat = pd.DataFrame({
'hospital_num': np.random.choice([f'P{i}' for i in range(1, 51)], 200),
'date_visit': np.random.choice(date_range, 200)
})
lab_dat = lab_dat.sort_values(['hospital_num', 'date_visit'])
# Sample groups
sampled_hospital_nums = np.random.choice(lab_dat['hospital_num'].unique(), 20, replace=False)
lab_dat = lab_dat[lab_dat['hospital_num'].isin(sampled_hospital_nums)]
# Calculate days since last test
lab_dat = lab_dat.sort_values(['hospital_num', 'date_visit'])
lab_dat['days_since_last_test'] = lab_dat.groupby('hospital_num')['date_visit'].diff().dt.days
# Select required columns
lab_dat = lab_dat[['hospital_num', 'days_since_last_test', 'date_visit']]
# Calculate longest test interval
longest_intervals = lab_dat.groupby('hospital_num')['days_since_last_test'].max()
lab_dat = lab_dat.merge(longest_intervals.rename('longest_test_interval'), on='hospital_num')
# Sort hospital numbers by longest test interval
hospital_num_order = lab_dat.groupby('hospital_num')['longest_test_interval'].max().sort_values().index
lab_dat['hospital_num'] = pd.Categorical(lab_dat['hospital_num'], categories=hospital_num_order, ordered=True)
# Plot
plt.figure(figsize=(12, 8))
for hospital in lab_dat['hospital_num'].unique():
hospital_data = lab_dat[lab_dat['hospital_num'] == hospital]
plt.plot(hospital_data['date_visit'], hospital_data['hospital_num'], marker='o')
plt.yticks(hospital_num_order)
plt.xlabel('Date Visit')
plt.ylabel('Hospital Number')
plt.title('Viral Load Tests')
plt.tight_layout()
plt.show()
```
```{python}
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from datetime import datetime, timedelta
# Generate 200 random records
start_date = datetime(2020, 1, 1)
end_date = datetime(2024, 8, 16)
date_range = pd.date_range(start=start_date, end=end_date)
lab_dat = (pd.DataFrame({
'hospital_num': np.random.choice([f'P{i}' for i in range(1, 51)], 200),
'date_visit': np.random.choice(date_range, 200)
})
.sort_values(['hospital_num', 'date_visit'])
.reset_index(drop=True))
viral_load_tests_plot = (
lab_dat
# Sample groups
.loc[lab_dat['hospital_num'].isin(np.random.choice(lab_dat['hospital_num'].unique(), 20, replace=False))]
# Calculate days since last test
.sort_values(['hospital_num', 'date_visit'])
.assign(days_since_last_test=lambda x: x.groupby('hospital_num')['date_visit'].diff().dt.days)
# Select required columns
[['hospital_num', 'days_since_last_test', 'date_visit']]
# Calculate longest test interval
.assign(longest_test_interval=lambda x: x.groupby('hospital_num')['days_since_last_test'].transform('max'))
# Sort hospital numbers by longest test interval
.assign(hospital_num=lambda x: pd.Categorical(
x['hospital_num'],
categories=x.groupby('hospital_num')['longest_test_interval'].max().sort_values().index,
ordered=True
))
)
# Plot
plt.figure(figsize=(12, 8))
for hospital in viral_load_tests_plot['hospital_num'].cat.categories:
hospital_data = viral_load_tests_plot[viral_load_tests_plot['hospital_num'] == hospital]
plt.plot(hospital_data['date_visit'], hospital_data['hospital_num'], marker='o')
plt.yticks(viral_load_tests_plot['hospital_num'].cat.categories)
plt.xlabel('Date Visit')
plt.ylabel('Hospital Number')
plt.title('Viral Load Tests')
plt.tight_layout()
plt.show()
```
```{python}
```