Exercise: Advanced Support Visualizations

Now push both tools beyond basic charts. The support data supports several advanced visualization types that communicate more information in less space. These are the visualizations that separate a data analyst from someone who can make bar charts.

Try each of these in both Claude and ChatGPT. Note which tool handles each request better.

1. HEATMAP: "Create a heatmap showing ticket count by category (rows) and priority (columns). Use a sequential color scale from dark to bright cyan. Include counts in each cell."

2. VIOLIN PLOT: "Create a violin plot of resolution hours, split by whether the ticket was escalated. Show the distribution shape, not just the median. Add individual data points as a strip plot overlay."

3. PARALLEL COORDINATES: "Create a parallel coordinates plot with axes for: Priority (ordinal), Resolution Hours, Satisfaction, and Escalated (binary). Color lines by category. This should reveal which categories cluster together across multiple dimensions."

4. SMALL MULTIPLES: "Create a 2x3 grid of small scatter plots — one per ticket category — showing resolution hours vs satisfaction. Same axis scales across all plots. Highlight escalated tickets with a different marker."

Claude will render these as Artifacts — you will see the charts immediately and can iterate. ChatGPT will produce Python scripts for each. For advanced chart types, the ChatGPT route has an advantage: matplotlib and seaborn give you precise control over violin plot bandwidths, parallel coordinate axis spacing, and strip plot jitter — details that Artifacts may not expose for fine-tuning.