visualization

Pack: 100xOS shared skills

Category: data-handling

Field: economics

License: private (curator-owned)

Updated: 2026-05-20

Stages: data-acquisition

Curator-private skill — copy text from 100xOS/shared/skills/data/visualization.md.

↗ view SKILL.md on source

Data Visualization Guide for Economics Research

Chart Selection by Data Relationship

Scatter Plot — Correlations and Cross-Sectional Relationships

Use when showing the relationship between two continuous variables across units.

Python

import matplotlib.pyplot as plt
import seaborn as sns

fig, ax = plt.subplots(figsize=(6, 5))
ax.scatter(df["log_gdp_pc"], df["life_expectancy"], s=20, alpha=0.6, edgecolors="none")
ax.set_xlabel("Log GDP per Capita")
ax.set_ylabel("Life Expectancy (years)")
ax.set_title("Income and Health, 2020")

## Add OLS fit line
sns.regplot(x="log_gdp_pc", y="life_expectancy", data=df, ax=ax,
            scatter=False, ci=95, line_kws={"color": "darkred", "linewidth": 1.5})

Enhancements: size points by population, color by region, add country labels for notable outliers.

Line Plot — Time Series

Use for showing trends, cycles, and changes over time.

Python

fig, ax = plt.subplots(figsize=(8, 4.5))
ax.plot(df["date"], df["gdp_growth"], linewidth=1.2, color="#1f77b4")
ax.axhline(y=0, color="gray", linewidth=0.5, linestyle="--")

## Shade recessions (NBER dates)
for start, end in recession_dates:
    ax.axvspan(start, end, alpha=0.15, color="gray")

ax.set_xlabel("")
ax.set_ylabel("Real GDP Growth (%)")
ax.set_title("U.S. Real GDP Growth, 1970-2024")

For multiple series, keep the number of lines to 4-5 maximum. Use distinct colors and add a legend outside the plot area if needed.

Bar Chart — Comparisons Across Categories

Use for comparing values across discrete groups.

Python

fig, ax = plt.subplots(figsize=(7, 5))
bars = ax.barh(df["country"], df["gini_coefficient"], color="#4c72b0", edgecolor="none")
ax.set_xlabel("Gini Coefficient")
ax.set_title("Income Inequality by Country, 2023")
ax.invert_yaxis()  # Highest value at top

Horizontal bars are preferred when category labels are long. Sort bars by value, not alphabetically, unless there is a natural ordering.

Coefficient Plot — Regression Results

Use instead of regression tables when the audience is broad or when comparing many specifications.

Python

import numpy as np

fig, ax = plt.subplots(figsize=(6, 5))

## Assuming results from statsmodels
coefs = results.params[1:]  # Exclude intercept
ci_low = results.conf_int()[0][1:]
ci_high = results.conf_int()[1][1:]
names = coefs.index

y_pos = np.arange(len(names))
ax.errorbar(coefs, y_pos, xerr=[coefs - ci_low, ci_high - coefs],
            fmt="o", color="#333333", ecolor="#999999", capsize=3, markersize=5)
ax.axvline(x=0, color="red", linewidth=0.8, linestyle="--")
ax.set_yticks(y_pos)
ax.set_yticklabels(names)
ax.set_xlabel("Coefficient Estimate (95% CI)")
ax.set_title("OLS Estimates")

Histogram / Density — Distributions

Use to show the shape of a variable's distribution.

Python

fig, ax = plt.subplots(figsize=(6, 4))
ax.hist(df["log_income"], bins=50, density=True, alpha=0.7, color="#4c72b0", edgecolor="white")
## Overlay kernel density
df["log_income"].plot.kde(ax=ax, color="darkred", linewidth=1.5)
ax.set_xlabel("Log Income")
ax.set_ylabel("Density")

Heatmap — Correlation Matrices and Two-Dimensional Summaries

Python

corr = df[vars_of_interest].corr()
fig, ax = plt.subplots(figsize=(8, 7))
sns.heatmap(corr, annot=True, fmt=".2f", cmap="RdBu_r", center=0,
            vmin=-1, vmax=1, square=True, ax=ax,
            linewidths=0.5, cbar_kws={"shrink": 0.8})

Binned Scatter Plot — Nonparametric Relationships

Common in applied economics (Chetty-style). Bins the x-variable and plots mean y within each bin.

Python

df["x_bin"] = pd.qcut(df["x"], q=20, duplicates="drop")
binned = df.groupby("x_bin").agg(x_mean=("x", "mean"), y_mean=("y", "mean")).reset_index()

fig, ax = plt.subplots(figsize=(6, 5))
ax.scatter(binned["x_mean"], binned["y_mean"], s=40, color="#333333", zorder=3)
ax.set_xlabel("X (binned means)")
ax.set_ylabel("Y (conditional mean)")

Accessibility — Colorblind-Safe Palettes

Approximately 8% of men and 0.5% of women have color vision deficiency. Always design for accessibility.

Recommended Palettes

Python

## Paul Tol's qualitative palette (up to 7 colors)
tol_bright = ["#4477AA", "#EE6677", "#228833", "#CCBB44", "#66CCEE", "#AA3377", "#BBBBBB"]

## IBM Design Library colorblind-safe
ibm_cb = ["#648FFF", "#785EF0", "#DC267F", "#FE6100", "#FFB000"]

## Seaborn colorblind palette
sns.set_palette("colorblind")

## For sequential data: viridis, cividis, or inferno (perceptually uniform)
plt.cm.viridis
plt.cm.cividis  # Best for colorblind readers

## For diverging data: RdBu (red-blue), PiYG (pink-green)

Additional Accessibility Guidelines

• Use shape and line style in addition to color to distinguish series.
• Ensure sufficient contrast (avoid light colors on white backgrounds).
• Add direct labels to lines rather than relying solely on legends.
• Use patterns or hatching in bar charts if color alone is insufficient.

Python

line_styles = ["-", "--", "-.", ":"]
markers = ["o", "s", "^", "D"]
for i, col in enumerate(columns):
    ax.plot(df["date"], df[col], linestyle=line_styles[i], marker=markers[i],
            markevery=10, label=col)

Publication-Quality Settings

Global Defaults

Python

import matplotlib as mpl

## Publication defaults
mpl.rcParams.update({
    "figure.figsize": (6.5, 4.5),       # Fits single-column journal width
    "figure.dpi": 150,                    # Screen display
    "savefig.dpi": 300,                   # Print quality
    "savefig.bbox": "tight",
    "savefig.transparent": False,
    "font.family": "serif",               # Or "sans-serif" for some journals
    "font.size": 10,
    "axes.titlesize": 11,
    "axes.labelsize": 10,
    "xtick.labelsize": 9,
    "ytick.labelsize": 9,
    "legend.fontsize": 9,
    "axes.linewidth": 0.8,
    "axes.spines.top": False,             # Remove top spine
    "axes.spines.right": False,           # Remove right spine
    "lines.linewidth": 1.2,
    "grid.alpha": 0.3,
    "grid.linewidth": 0.5,
})

Saving for Publication

Python

## Vector formats for journals (scalable, small file size for line art)
fig.savefig("figure1.pdf", format="pdf")
fig.savefig("figure1.eps", format="eps")

## Raster for presentations or web
fig.savefig("figure1.png", format="png", dpi=300)

## For LaTeX integration
fig.savefig("figure1.pgf")  # Native LaTeX rendering of text

Common Journal Requirements

Journal Type	Width (inches)	Font	Format
Single column	3.25-3.5	Matching journal	PDF/EPS
Double column	6.5-7.0	Matching journal	PDF/EPS
AER / QJE / RES	6.5	Times/Computer Modern	PDF
Presentation slides	10 x 6	Sans-serif	PNG (high DPI)

Multi-Panel Figures

Python

fig, axes = plt.subplots(2, 2, figsize=(7, 6), constrained_layout=True)

for i, (ax, var) in enumerate(zip(axes.flat, variables)):
    ax.plot(df["date"], df[var])
    ax.set_title(f"({chr(97+i)}) {var_labels[var]}")  # (a), (b), (c), (d)

fig.savefig("figure_panels.pdf")

Annotations and Notes

Python

## Source note below figure
fig.text(0.01, -0.02, "Source: FRED, Federal Reserve Bank of St. Louis.",
         fontsize=8, color="gray", ha="left", transform=fig.transFigure)

## Note about methodology
fig.text(0.01, -0.05, "Note: Shaded areas indicate NBER recession dates.",
         fontsize=8, color="gray", ha="left", transform=fig.transFigure)

Quick Reference — Choosing the Right Chart

Question	Chart Type
How does Y relate to X?	Scatter
How does Y change over time?	Line
How does Y differ across groups?	Bar (horizontal)
What is the distribution of Y?	Histogram / density
What are my regression estimates?	Coefficient plot
How do many variables correlate?	Heatmap
What is the nonparametric shape?	Binned scatter
How do two distributions compare?	Overlaid densities
What is the composition over time?	Stacked area
Geographic variation?	Choropleth map