rddensity/Python/rddensity_plot_illustration.py at main · rdpackages/rddensity · GitHub

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
###########################################################################
## RDDENSITY Python Package
## Script for RDDENSITY Plot Illustration
###########################################################################

from pathlib import Path
import math

import numpy as np
import pandas as pd
import plotnine as pn
from lpdensity import lpdensity
from scipy.stats import norm

import rddensity


### Load rddensity package
### NOTE: depending on your system, you may need to install it first
# pip install rddensity

### Load data base
data = pd.read_csv(Path(__file__).resolve().with_name("rddensity_senate.csv"))
margin = data["margin"]

### Plot and save results
rdd = rddensity.rddensity(X=margin)
rdplot = rddensity.rdplotdensity(rdd, X=margin)

### Construct the same plot
c = rdd.c
hl = rdd.h["left"]
hr = rdd.h["right"]
plot_range = [max(margin.min(), c - 3 * hl), min(margin.max(), c + 3 * hr)]
plot_n = [10, 10]

grid_left = np.linspace(plot_range[0], c, plot_n[0] + 1)
grid_left[plot_n[0]] = c
grid_right = np.linspace(c, plot_range[1], plot_n[1] + 1)
grid_right[0] = c

X = pd.DataFrame({"margin": margin}).dropna()
X_left = X[X["margin"] < c]
X_right = X[X["margin"] >= c]
scale_left = [(len(X_left) - 1) / (len(X) - 1)]
scale_right = [(len(X_right) - 1) / (len(X) - 1)]

Estl = lpdensity(
    data=X_left,
    grid=grid_left,
    bw=[hl],
    p=rdd.p,
    q=rdd.q,
    v=1,
    kernel=rdd.kernel,
    scale=scale_left,
    regularize=rdd.regularize,
    nLocalMin=rdd.nLocalMin,
    nUniqueMin=rdd.nUniqueMin,
    massPoints=rdd.massPoints,
)
Estr = lpdensity(
    data=X_right,
    grid=grid_right,
    bw=[hr],
    p=rdd.p,
    q=rdd.q,
    v=1,
    kernel=rdd.kernel,
    scale=scale_right,
    regularize=rdd.regularize,
    nLocalMin=rdd.nLocalMin,
    nUniqueMin=rdd.nUniqueMin,
    massPoints=rdd.massPoints,
)

### CI: below the cutoff
rdplot_left = Estl.Estimate.copy()
rdplot_left["cil"] = rdplot_left["f_q"] - norm.ppf(0.975) * rdplot_left["se_q"]
rdplot_left["ciu"] = rdplot_left["f_q"] + norm.ppf(0.975) * rdplot_left["se_q"]

### CI: above the cutoff
rdplot_right = Estr.Estimate.copy()
rdplot_right["cil"] = rdplot_right["f_q"] - norm.ppf(0.975) * rdplot_right["se_q"]
rdplot_right["ciu"] = rdplot_right["f_q"] + norm.ppf(0.975) * rdplot_right["se_q"]

### Histogram
N_left = int(((margin >= rdplot_left["grid"].min()) & (margin < c)).sum())
hist_n_left = math.ceil(min(math.sqrt(N_left), 10 * math.log(N_left) / math.log(10)))

N_right = int(((margin <= rdplot_right["grid"].max()) & (margin >= c)).sum())
hist_n_right = math.ceil(min(math.sqrt(N_right), 10 * math.log(N_right) / math.log(10)))

hist_breaks = np.concatenate(
    [
        np.linspace(rdplot_left["grid"].min(), c, hist_n_left + 1),
        np.linspace(c, rdplot_right["grid"].max(), hist_n_right + 1)[1:],
    ]
)
hist_scale = ((margin >= rdplot_left["grid"].min()) & (margin <= rdplot_right["grid"].max())).mean()
hist_data = pd.DataFrame({"margin": margin})

### Call plotnine
manual_plot = (
    pn.ggplot()
    + pn.theme_bw()
    + pn.geom_histogram(
        data=hist_data,
        mapping=pn.aes(x="margin", y=pn.after_stat(f"density * {hist_scale:.17g}")),
        breaks=hist_breaks,
        fill="#00BA38",
        color="white",
        alpha=0.2,
    )
    + pn.geom_ribbon(
        data=rdplot_left,
        mapping=pn.aes(x="grid", ymin="cil", ymax="ciu"),
        alpha=0.2,
        fill="black",
    )
    + pn.geom_ribbon(
        data=rdplot_right,
        mapping=pn.aes(x="grid", ymin="cil", ymax="ciu"),
        alpha=0.2,
        fill="red",
    )
    + pn.geom_line(data=rdplot_left, mapping=pn.aes(x="grid", y="f_p"), color="black")
    + pn.geom_line(data=rdplot_right, mapping=pn.aes(x="grid", y="f_p"), color="red")
    + pn.labs(x="margin", y="")
)