Skip to content

Commit 0037242

Browse files
github-actions[bot]github-actions[bot]
authored
feat(seaborn): implement forest-basic (#2395)
## Implementation: `forest-basic` - seaborn Implements the **seaborn** version of `forest-basic`. **File:** `plots/forest-basic/implementations/seaborn.py` --- :robot: *[impl-generate workflow](https://github.com/MarkusNeusinger/pyplots/actions/runs/20543281061)* --------- Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com> Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
1 parent 100919c commit 0037242

2 files changed

Lines changed: 195 additions & 0 deletions

File tree

‎plots/forest-basic/implementations/seaborn.py‎

Lines changed: 168 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,168 @@
1+
""" pyplots.ai
2+
forest-basic: Meta-Analysis Forest Plot
3+
Library: seaborn 0.13.2 | Python 3.13.11
4+
Quality: 92/100 | Created: 2025-12-27
5+
"""
6+
7+
import matplotlib.patches as mpatches
8+
import matplotlib.pyplot as plt
9+
import numpy as np
10+
import pandas as pd
11+
import seaborn as sns
12+
13+
14+
# Data: Meta-analysis of treatment effect (mean difference) from 10 studies
15+
np.random.seed(42)
16+
17+
studies = [
18+
"Smith et al. 2018",
19+
"Johnson et al. 2019",
20+
"Williams et al. 2019",
21+
"Brown et al. 2020",
22+
"Davis et al. 2020",
23+
"Miller et al. 2021",
24+
"Wilson et al. 2021",
25+
"Moore et al. 2022",
26+
"Taylor et al. 2022",
27+
"Anderson et al. 2023",
28+
]
29+
30+
# Effect sizes (mean differences) - some favor treatment, some favor control
31+
effect_sizes = [-0.45, 0.12, -0.28, -0.52, 0.05, -0.38, -0.15, -0.42, -0.22, -0.35]
32+
ci_widths = [0.35, 0.28, 0.42, 0.25, 0.55, 0.32, 0.38, 0.30, 0.45, 0.28]
33+
ci_lower = [e - w for e, w in zip(effect_sizes, ci_widths, strict=True)]
34+
ci_upper = [e + w for e, w in zip(effect_sizes, ci_widths, strict=True)]
35+
weights = [12.5, 8.2, 6.8, 14.1, 5.5, 10.3, 7.9, 11.8, 6.2, 9.7]
36+
37+
# Calculate pooled estimate (weighted mean)
38+
pooled_effect = np.average(effect_sizes, weights=weights)
39+
pooled_se = np.sqrt(1 / np.sum([w / (ci_w**2) for w, ci_w in zip(weights, ci_widths, strict=True)]))
40+
pooled_ci_lower = pooled_effect - 1.96 * pooled_se
41+
pooled_ci_upper = pooled_effect + 1.96 * pooled_se
42+
43+
df = pd.DataFrame(
44+
{"study": studies, "effect": effect_sizes, "ci_lower": ci_lower, "ci_upper": ci_upper, "weight": weights}
45+
)
46+
47+
# Sort by effect size
48+
df = df.sort_values("effect", ascending=True).reset_index(drop=True)
49+
50+
# Create figure
51+
fig, ax = plt.subplots(figsize=(16, 9))
52+
sns.set_style("whitegrid")
53+
54+
# Y positions for studies (leave space at bottom for pooled estimate)
55+
y_positions = np.arange(len(df)) + 1.5
56+
57+
# Scale marker sizes based on weight (larger for more precise studies)
58+
marker_sizes = (df["weight"] / df["weight"].max()) * 400 + 100
59+
60+
# Plot confidence intervals as horizontal lines
61+
for i, (_, row) in enumerate(df.iterrows()):
62+
ax.hlines(y=y_positions[i], xmin=row["ci_lower"], xmax=row["ci_upper"], color="#306998", linewidth=2.5, zorder=1)
63+
64+
# Plot point estimates using seaborn
65+
sns.scatterplot(
66+
data=df,
67+
x="effect",
68+
y=y_positions,
69+
size="weight",
70+
sizes=(100, 500),
71+
color="#306998",
72+
edgecolor="white",
73+
linewidth=1.5,
74+
legend=False,
75+
ax=ax,
76+
zorder=2,
77+
)
78+
79+
# Add study labels on the left
80+
for i, (_, row) in enumerate(df.iterrows()):
81+
ax.text(-1.4, y_positions[i], row["study"], fontsize=14, va="center", ha="left", fontweight="medium")
82+
83+
# Add effect size values on the right
84+
for i, (_, row) in enumerate(df.iterrows()):
85+
ax.text(
86+
1.1,
87+
y_positions[i],
88+
f"{row['effect']:.2f} [{row['ci_lower']:.2f}, {row['ci_upper']:.2f}]",
89+
fontsize=12,
90+
va="center",
91+
ha="left",
92+
family="monospace",
93+
)
94+
95+
# Draw pooled estimate diamond
96+
diamond_y = 0.3
97+
diamond_height = 0.4
98+
diamond = mpatches.Polygon(
99+
[
100+
[pooled_effect, diamond_y],
101+
[pooled_ci_lower, diamond_y + diamond_height / 2],
102+
[pooled_effect, diamond_y + diamond_height],
103+
[pooled_ci_upper, diamond_y + diamond_height / 2],
104+
],
105+
closed=True,
106+
facecolor="#FFD43B",
107+
edgecolor="#306998",
108+
linewidth=2,
109+
zorder=3,
110+
)
111+
ax.add_patch(diamond)
112+
113+
# Add pooled estimate label
114+
ax.text(-1.4, diamond_y + diamond_height / 2, "Pooled Estimate", fontsize=14, va="center", ha="left", fontweight="bold")
115+
ax.text(
116+
1.1,
117+
diamond_y + diamond_height / 2,
118+
f"{pooled_effect:.2f} [{pooled_ci_lower:.2f}, {pooled_ci_upper:.2f}]",
119+
fontsize=12,
120+
va="center",
121+
ha="left",
122+
family="monospace",
123+
fontweight="bold",
124+
)
125+
126+
# Vertical reference line at null effect (0)
127+
ax.axvline(x=0, color="#666666", linestyle="--", linewidth=2, zorder=0, alpha=0.7)
128+
129+
# Separator line above pooled estimate
130+
ax.axhline(y=1.0, color="#CCCCCC", linewidth=1.5, zorder=0)
131+
132+
# Styling
133+
ax.set_xlim(-1.5, 1.8)
134+
ax.set_ylim(-0.3, len(df) + 2)
135+
ax.set_xlabel("Mean Difference (Treatment - Control)", fontsize=20)
136+
ax.set_ylabel("")
137+
ax.set_title("forest-basic · seaborn · pyplots.ai", fontsize=24, fontweight="bold", pad=20)
138+
139+
# Remove y-axis ticks (study names are shown as text)
140+
ax.set_yticks([])
141+
142+
# Style x-axis ticks
143+
ax.tick_params(axis="x", labelsize=16)
144+
145+
# Add annotation for interpretation
146+
ax.text(
147+
-0.75,
148+
len(df) + 1.5,
149+
"← Favors Treatment",
150+
fontsize=14,
151+
ha="center",
152+
va="center",
153+
color="#306998",
154+
fontweight="medium",
155+
)
156+
ax.text(
157+
0.75, len(df) + 1.5, "Favors Control →", fontsize=14, ha="center", va="center", color="#666666", fontweight="medium"
158+
)
159+
160+
# Adjust grid
161+
ax.grid(axis="x", alpha=0.3, linestyle="--")
162+
ax.grid(axis="y", visible=False)
163+
164+
# Remove top and right spines
165+
sns.despine(left=True)
166+
167+
plt.tight_layout()
168+
plt.savefig("plot.png", dpi=300, bbox_inches="tight")

‎plots/forest-basic/metadata/seaborn.yaml‎

Lines changed: 27 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,27 @@
1+
library: seaborn
2+
specification_id: forest-basic
3+
created: '2025-12-27T19:21:16Z'
4+
updated: '2025-12-27T19:28:36Z'
5+
generated_by: claude-opus-4-5-20251101
6+
workflow_run: 20543281061
7+
issue: 0
8+
python_version: 3.13.11
9+
library_version: 0.13.2
10+
preview_url: https://storage.googleapis.com/pyplots-images/plots/forest-basic/seaborn/plot.png
11+
preview_thumb: https://storage.googleapis.com/pyplots-images/plots/forest-basic/seaborn/plot_thumb.png
12+
preview_html: null
13+
quality_score: 92
14+
review:
15+
strengths:
16+
- 'Excellent implementation of all forest plot requirements: diamond pooled estimate,
17+
weighted markers, reference line, CI whiskers'
18+
- Clear, professional layout with study labels and effect sizes displayed in readable
19+
format
20+
- Good use of seaborn scatterplot for weighted point estimates with size mapping
21+
- Proper directional annotations help interpret the plot
22+
- Realistic meta-analysis data with chronological study names and varied effect
23+
sizes
24+
weaknesses:
25+
- sns.set_style called after plt.subplots (should be called before for idiomatic
26+
seaborn usage)
27+
- Layout could be slightly tighter on the right side to reduce whitespace

0 commit comments

Comments
 (0)