Key Takeaways: Communicating with Data: Telling Stories with Numbers

One-Sentence Summary

Effective data communication requires honest visualizations that follow Tufte's principles (maximize data-ink, minimize chartjunk), audience-appropriate writing that includes both statistical significance and effect sizes, structured reports with Introduction-Methods-Results-Discussion-Limitations, and reproducible analysis practices — because the most rigorous analysis in the world is worthless if your audience doesn't understand it.

Core Concepts at a Glance

Tufte's Principles

Misleading Techniques Checklist

Writing Statistical Results

Template Sentences

Confidence Interval:

Technical: "The 95% CI for mean [variable] was ([lower], [upper])."

Plain: "We estimate the average [variable] is between [lower] and [upper]."

t-Test:

Technical: "t([df]) = [value], p = [value], d = [value], 95% CI: ([lower], [upper])."

Plain: "[Group 1] scored [higher/lower] by about [difference]. This is [unlikely] to be chance, and the effect is [small/medium/large]."

Regression:

Technical: "b = [slope], p = [value], $R^2$ = [value]."

Plain: "For every additional [unit of x], [y] tends to [change] by about [slope]. The model explains [R² × 100]% of the variation."

The "So What?" Checklist

Every result needs: 1. The finding — what happened 2. The magnitude — how big (effect size) 3. The uncertainty — how confident (CI) 4. The implication — so what? (recommendation)

Report Structure

The Five Sections

Executive Summary Template

1. What did we study? (One sentence)
2. What did we find? (One or two sentences)
3. Why does it matter? (One sentence)
4. What should we do? (One sentence)

Presenting Uncertainty Honestly

Hedging Language Guide

Accessibility Principles

Reproducibility Checklist

Key Python Code

Professional Chart Template

import matplotlib.pyplot as plt
import seaborn as sns
import numpy as np

# Set global style
sns.set_style("whitegrid")
plt.rcParams.update({
    'font.size': 11,
    'axes.titlesize': 14,
    'figure.dpi': 150,
    'savefig.dpi': 300
})

fig, ax = plt.subplots(figsize=(8, 5))

# Plot data
ax.bar(categories, values, color='steelblue', edgecolor='none')

# Clean styling
ax.spines['top'].set_visible(False)
ax.spines['right'].set_visible(False)
ax.spines['left'].set_color('#CCCCCC')
ax.spines['bottom'].set_color('#CCCCCC')
ax.grid(axis='y', alpha=0.3)

# Descriptive title (states the finding)
ax.set_title('Monthly Sales Increased 12% After Campaign',
             color='#333333')
ax.set_ylabel('Sales ($K)', color='#555555')

# Data labels
for bar, val in zip(ax.patches, values):
    ax.text(bar.get_x() + bar.get_width() / 2,
            bar.get_height() + 1, f'${val}K',
            ha='center', fontsize=10)

plt.tight_layout()
plt.savefig('chart.png', dpi=300, bbox_inches='tight')

Annotation Example

# Add annotation to highlight key data point
ax.annotate('Peak: 380 visits',
            xy=(x_val, y_val),
            xytext=(x_offset, y_offset),
            arrowprops=dict(arrowstyle='->',
                           color='#E74C3C'),
            fontsize=10, color='#E74C3C',
            fontweight='bold')

# Add reference line
ax.axhline(y=target, color='gray', linestyle='--',
           linewidth=1, alpha=0.5)
ax.text(x_pos, target - 3, 'Target: 120 min',
        fontsize=9, color='gray')

Excel Chart Formatting

Common Mistakes

Connections