Day 5: Building a Complete Data Story

Narrative Visualization Principles

Data without narrative is just numbers. The best data visualizations answer: what does this mean, why does it matter, and what should we do?

Three principles for narrative visualization:

• Lead with the conclusion. Tell viewers what they're about to see before they see it. "Revenue grew 40% — here's what drove it."
• Each chart answers one question. If you need a chart to do too much, split it into two.
• Annotations carry the story. A trend line without annotation is a shape. With annotation ("Q3 surge driven by enterprise deal") it's insight.

The Four-Chart Story Structure

Most business data stories need four charts in sequence:

Chart 1: The overview
  Show the main metric over time. Set context.
  Example: Monthly revenue, 18-month trend

Chart 2: The composition
  Break it down. Show what makes up the number.
  Example: Revenue by product/region/channel

Chart 3: The anomaly or insight
  Highlight the interesting pattern or pivot point.
  Example: One segment growing while others flat

Chart 4: The implication
  Show what it means going forward.
  Example: Projected growth if trend continues vs. if it doesn't

Build a Complete Data Story

Build a Streamlit app that tells a complete story with real or sample data. Use this structure:

import streamlit as st
import plotly.express as px

st.title("Q3 2025 Business Review")
st.markdown("The bottom line: Revenue grew 22% YoY.")

st.subheader("1. Overall growth is strong")
# [line chart code here]

st.subheader("2. Enterprise segment is doing the work")
# [bar chart breakdown here]

st.subheader("3. SMB churn is the risk to watch")
# [trend chart highlighting divergence]

st.markdown("### What this means: Q4 target achievable if trends hold.")

Notice: every chart section has a narrative header that states the conclusion, not just the metric. The story structure means a reader who only reads the headers knows everything important — the charts are supporting evidence.