Tim Diller
b75a427aad
goal is to leave lots of room for student improvements in class. also, introduce missingness in the data
47 lines
1.2 KiB
Python
47 lines
1.2 KiB
Python
import random
|
|
import sys
|
|
|
|
import numpy as np
|
|
|
|
|
|
FILE_COLUMN_NAMES = [
|
|
"n", "m", "v1", "v2", "v3", "p1", "p2", "p3",
|
|
]
|
|
|
|
|
|
def write_data_file(filename, n=20, missing_values=False):
|
|
"""Write a csv file with data
|
|
"""
|
|
|
|
with open(filename, "w") as fp:
|
|
fp.write(", ".join(FILE_COLUMN_NAMES) + ",\n")
|
|
for i in range(n):
|
|
vals = []
|
|
vals.append(random.gauss(3., 0.5))
|
|
vals.extend(
|
|
(
|
|
random.gauss(0., 25),
|
|
random.gauss(0., 25),
|
|
random.gauss(0., 25),
|
|
)
|
|
)
|
|
vals.extend(
|
|
(
|
|
random.gauss(0., 1.5),
|
|
random.gauss(0., 1.5),
|
|
random.gauss(0., 1.5),
|
|
)
|
|
)
|
|
if missing_values:
|
|
if random.randint(0,10) < 1:
|
|
idx = random.randint(1, len(vals))
|
|
vals[idx] = np.nan
|
|
line = f"{i}, " + ", ".join(f"{v:7.5f}" for v in vals) + ",\n"
|
|
fp.write(line)
|
|
|
|
|
|
if __name__ == '__main__':
|
|
missing_values = sys.argv[-1] == '-m'
|
|
|
|
write_data_file("sample_data.csv", n=20, missing_values=True)
|