From 5bfc3d9600319bac9070c816c9a43e6e30e0e255 Mon Sep 17 00:00:00 2001
From: medusa <newton214@gmail.com>
Date: Wed, 2 Jul 2025 07:02:31 +0000
Subject: [PATCH] Update tech_docs/webdev/hugo.md

---
 tech_docs/webdev/hugo.md | 119 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 119 insertions(+)

diff --git a/tech_docs/webdev/hugo.md b/tech_docs/webdev/hugo.md
index 0e1ea59..2ef3a89 100644
--- a/tech_docs/webdev/hugo.md
+++ b/tech_docs/webdev/hugo.md
@@ -1,3 +1,122 @@
+For **statistics and data science visuals** in hobby projects (e.g., distributions, regression plots, time series, clustering), here’s the optimal workflow balancing **quality, performance, and ease of use**:
+
+---
+
+### **1. Recommended Tools by Task**
+#### **A. Static Visuals (PNG/SVG)**
+| Task                  | Best Tool (Python)       | Best Tool (R)          | Output Format |
+|-----------------------|--------------------------|------------------------|---------------|
+| Distributions         | Seaborn (`histplot`, `kdeplot`) | ggplot2 (`geom_histogram`, `geom_density`) | PNG (if dense data) / SVG (if simple) |
+| Scatter/Regression    | Seaborn (`regplot`, `lmplot`) | ggplot2 (`geom_smooth`, `geom_point`) | SVG (for interactivity later) |
+| Time Series          | Matplotlib (`plot_date`) + Seaborn | ggplot2 (`geom_line`) | PNG (long series) |
+| Heatmaps/Correlation | Seaborn (`heatmap`)       | ggplot2 (`geom_tile`)  | PNG (large matrices) |
+| Box/Violin Plots     | Seaborn (`boxplot`, `violinplot`) | ggplot2 (`geom_boxplot`) | SVG |
+
+#### **B. Interactive Visuals (Embeddable in Hugo)**
+| Task                  | Best Tool                | Output Format          |
+|-----------------------|--------------------------|------------------------|
+| Exploratory hover plots | Plotly Express (Python/R) | HTML (embed as `<iframe>`) |
+| Linked brushing       | Altair (Python) + Vega   | JSON (via Hugo shortcode) |
+| Animated transitions  | Plotly (`animation_frame`) | GIF/MP4 (FFmpeg-optimized) |
+
+---
+
+### **2. File Format Guidelines**
+- **Use PNG if:**  
+  - Plot has >1,000 data points (e.g., large scatterplots).  
+  - You’re using transparency (`plt.savefig(..., transparent=True)`).  
+  - Example:  
+    ```python
+    # Python (Seaborn)
+    sns.scatterplot(data=df, x="x", y="y").get_figure().savefig("plot.png", dpi=200, bbox_inches="tight")
+    ```
+
+- **Use SVG if:**  
+  - Plot is simple (e.g., boxplots, small line charts).  
+  - You want to edit colors/text later in Inkscape.  
+  - Example:  
+    ```r
+    # R (ggplot2)
+    ggsave("plot.svg", plot, device = svg, width=8, height=6)
+    ```
+
+---
+
+### **3. Optimization Pipeline**
+#### **For PNGs:**
+1. Generate high-res (300dpi) PNGs directly from Python/R.  
+2. Compress with `oxipng` (lossless):  
+   ```bash
+   oxipng -o 4 --strip safe plot.png
+   ```
+   - (Alternative: Use ImageMagick `-quality 90 -strip`).  
+
+#### **For SVGs:**
+1. Export from tool (avoid unnecessary metadata).  
+2. Crush with `svgo`:  
+   ```bash
+   npx svgo plot.svg --config='
+   {
+     "plugins": [
+       "removeXMLProcInst",
+       "removeComments",
+       "removeDoctype"
+     ]
+   }'
+   ```
+
+---
+
+### **4. Hugo Integration**
+- **Static images**: Place in `/static/images/` and reference in Markdown:  
+  ```md
+  ![Distribution Plot](/images/distplot.png)
+  ```
+- **Interactive plots**: Use Hugo shortcodes for iframes:  
+  ```html
+  {{< iframe src="/plots/interactive_plot.html" width="100%" height="500px" >}}
+  ```
+
+---
+
+### **5. Example Workflows**
+#### **Python (Seaborn + Plotly)**
+```python
+# Static plot
+sns.histplot(data=df, x="value").get_figure().savefig("histogram.png", dpi=300)
+
+# Interactive plot
+fig = px.scatter(df, x="x", y="y", hover_data=["label"])
+fig.write_html("scatter.html")  # Embed in Hugo
+```
+
+#### **R (ggplot2 + Plotly)**
+```r
+# Static plot
+ggsave("boxplot.svg", ggplot(df, aes(x=group, y=value)) + geom_boxplot())
+
+# Interactive plot
+ggplotly(p) %>% htmlwidgets::saveWidget("boxplot.html")
+```
+
+---
+
+### **6. Tools to Avoid**
+- **D3.js**: Overkill unless you need custom interactivity.  
+- **Bokeh**: Heavy for static sites; Plotly is simpler.  
+- **RAWGraphs**: Better for one-offs, not automation.  
+
+---
+
+### **Final Recommendation**
+- **Default to Seaborn (Python) / ggplot2 (R)** → Export as **PNG** for dense plots, **SVG** for simple ones.  
+- **Add interactivity sparingly** with Plotly (saved as HTML).  
+- **Optimize all files** (`oxipng`/`svgo`) before Hugo deployment.  
+
+For hobby projects, this balances **speed, quality, and fun**! Want a sample Hugo project structure with automation scripts?
+
+---
+
 To get started with your data visualization project using Hugo, Bootstrap, and Chart.js, follow these steps:
 
 1. Project Setup: