Let the data speak first

EDA is where you build intuition before modeling. You ask questions, make plots, and let patterns — and problems — surface.

What to look for

• Distributions — histograms of each key variable.
• Relationships — scatter plots, correlations.
• Groups — how do segments differ?
• Anomalies — surprising spikes, gaps, or clusters.

df.hist(figsize=(10,6))
df.corr(numeric_only=True)
df.groupby("segment")["spend"].describe()

Takeaway: good EDA turns a spreadsheet into a story — and tells you what to model (and what to fix) next.