Create Viz

1---2name: create-viz3description: Create publication-quality visualizations with Python. Use when turning query results or a DataFrame into a chart, selecting the right chart type for a trend or comparison, generating a plot for a report or presentation, or needing an interactive chart with hover and zoom.4argument-hint: "<data source> [chart type]"5---6 7# /create-viz - Create Visualizations8 9> If you see unfamiliar placeholders or need to check which tools are connected, see [CONNECTORS.md](../../CONNECTORS.md).10 11Create publication-quality data visualizations using Python. Generates charts from data with best practices for clarity, accuracy, and design.12 13## Usage14 15```16/create-viz <data source> [chart type] [additional instructions]17```18 19## Workflow20 21### 1. Understand the Request22 23Determine:24 25- **Data source**: Query results, pasted data, CSV/Excel file, or data to be queried26- **Chart type**: Explicitly requested or needs to be recommended27- **Purpose**: Exploration, presentation, report, dashboard component28- **Audience**: Technical team, executives, external stakeholders29 30### 2. Get the Data31 32**If data warehouse is connected and data needs querying:**331. Write and execute the query342. Load results into a pandas DataFrame35 36**If data is pasted or uploaded:**371. Parse the data into a pandas DataFrame382. Clean and prepare as needed (type conversions, null handling)39 40**If data is from a previous analysis in the conversation:**411. Reference the existing data42 43### 3. Select Chart Type44 45If the user didn't specify a chart type, recommend one based on the data and question:46 47| Data Relationship | Recommended Chart |48|---|---|49| Trend over time | Line chart |50| Comparison across categories | Bar chart (horizontal if many categories) |51| Part-to-whole composition | Stacked bar or area chart (avoid pie charts unless <6 categories) |52| Distribution of values | Histogram or box plot |53| Correlation between two variables | Scatter plot |54| Two-variable comparison over time | Dual-axis line or grouped bar |55| Geographic data | Choropleth map |56| Ranking | Horizontal bar chart |57| Flow or process | Sankey diagram |58| Matrix of relationships | Heatmap |59 60Explain the recommendation briefly if the user didn't specify.61 62### 4. Generate the Visualization63 64Write Python code using one of these libraries based on the need:65 66- **matplotlib + seaborn**: Best for static, publication-quality charts. Default choice.67- **plotly**: Best for interactive charts or when the user requests interactivity.68 69**Code requirements:**70 71```python72import matplotlib.pyplot as plt73import seaborn as sns74import pandas as pd75 76# Set professional style77plt.style.use('seaborn-v0_8-whitegrid')78sns.set_palette("husl")79 80# Create figure with appropriate size81fig, ax = plt.subplots(figsize=(10, 6))82 83# [chart-specific code]84 85# Always include:86ax.set_title('Clear, Descriptive Title', fontsize=14, fontweight='bold')87ax.set_xlabel('X-Axis Label', fontsize=11)88ax.set_ylabel('Y-Axis Label', fontsize=11)89 90# Format numbers appropriately91# - Percentages: '45.2%' not '0.452'92# - Currency: '$1.2M' not '1200000'93# - Large numbers: '2.3K' or '1.5M' not '2300' or '1500000'94 95# Remove chart junk96ax.spines['top'].set_visible(False)97ax.spines['right'].set_visible(False)98 99plt.tight_layout()100plt.savefig('chart_name.png', dpi=150, bbox_inches='tight')101plt.show()102```103 104### 5. Apply Design Best Practices105 106**Color:**107- Use a consistent, colorblind-friendly palette108- Use color meaningfully (not decoratively)109- Highlight the key data point or trend with a contrasting color110- Grey out less important reference data111 112**Typography:**113- Descriptive title that states the insight, not just the metric (e.g., "Revenue grew 23% YoY" not "Revenue by Month")114- Readable axis labels (not rotated 90 degrees if avoidable)115- Data labels on key points when they add clarity116 117**Layout:**118- Appropriate whitespace and margins119- Legend placement that doesn't obscure data120- Sorted categories by value (not alphabetically) unless there's a natural order121 122**Accuracy:**123- Y-axis starts at zero for bar charts124- No misleading axis breaks without clear notation125- Consistent scales when comparing panels126- Appropriate precision (don't show 10 decimal places)127 128### 6. Save and Present129 1301. Save the chart as a PNG file with descriptive name1312. Display the chart to the user1323. Provide the code used so they can modify it1334. Suggest variations (different chart type, different grouping, zoomed time range)134 135## Examples136 137```138/create-viz Show monthly revenue for the last 12 months as a line chart with the trend highlighted139```140 141```142/create-viz Here's our NPS data by product: [pastes data]. Create a horizontal bar chart ranking products by score.143```144 145```146/create-viz Query the orders table and create a heatmap of order volume by day-of-week and hour147```148 149## Tips150 151- If you want interactive charts (hover, zoom, filter), mention "interactive" and Claude will use plotly152- Specify "presentation" if you need larger fonts and higher contrast153- You can request multiple charts at once (e.g., "create a 2x2 grid of charts showing...")154- Charts are saved to your current directory as PNG files