Simulating data in Python

Which environment?

import sys

def get_env():
    sp = sys.path[1].split("/")
    if "envs" in sp:
        return sp[sp.index("envs") + 1]
    else:
        return ""

get_env()

'fast_ai_course_2022'

Importing the required libraries

import numpy as np
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
from patsy import dmatrices, dmatrix

Simulating the data

n = 100
sd = 2

dat = pd.DataFrame()
dat["x1"] = np.random.normal(0, 5, n)
dat["x2"] = np.random.normal(0, 5, n)
dat["x3"] = np.random.choice(["A", "B"], size=n, replace=True)
dat["epsilon"] = np.random.normal(0, sd, n)
X = dmatrix(" ~ x1 + x2 + x3", dat)
X.design_info.column_names
beta = [14, 50, 1, -2]

dat["y"] = np.matmul(X, beta) + dat["epsilon"]
dat

	x1	x2	x3	epsilon	y
0	-7.579337	-1.600375	B	1.394259	61.015673
1	1.827419	1.912467	B	-2.066038	59.936446
2	0.916864	3.079072	B	-2.647573	56.111148
3	-8.580745	11.137583	B	-0.409475	32.734615
4	11.100234	4.150596	B	0.588830	67.387871
...	...	...	...	...	...
95	-5.862918	-7.058553	B	-1.026161	71.228028
96	3.791414	2.979757	A	-2.858461	8.973439
97	4.200092	3.668948	B	-1.505371	59.356825
98	-0.381726	-1.413223	B	2.492208	68.936928
99	-6.374441	-0.552026	A	3.637649	12.367261

100 rows × 5 columns

Showing the data

sns.scatterplot(data=dat, x="x1", y="y");

sns.scatterplot(data=dat, x="x2", y="y");

sns.scatterplot(data=dat, x="x2", y="y", hue="x3");