Add ggplot contact matrix heatmaps to explain symmetry in contact-matrices episode

[avallecam]

Context

The contact-matrices episode explains symmetric contact matrices but currently lacks a visual comparison that makes the symmetry argument concrete.

The script 73-socialmixr-ggplot.R in epicatador has two helper functions (format_big_mark, matrix_to_ggplot) that produce ggplot heatmaps of the contact matrix.

Plan

Add a single chunk (on a new branch) that:

1. Defines format_big_mark() — formats tile labels as human-readable numbers (K/M).
2. Defines matrix_to_ggplot() — converts a contact matrix to a geom_tile heatmap using colorspace::scale_fill_continuous_sequential(palette = "OrYel").
3. Produces two plots side by side or sequentially:
   • Plot 1: contacts_byage$matrix — mean contacts (asymmetric rates).
   • Plot 2: contacts_byage$matrix * contacts_byage$demography$population — total contacts (symmetric).

The contrast between the two plots makes the symmetry argument visual and explicit.

Adaptations from source

• Uses standard contacts_byage object (already loaded upstream via contactsurveys pattern).
• Uses |> instead of %>%.
• Explicit pkg:: namespacing.
• No socialmixr::get_survey() — survey loaded upstream.

Dependencies

• colorspace (for scale_fill_continuous_sequential)
• scales (for label_number)
• ggplot2, dplyr (already used in the episode)

[avallecam]

this will be one spoiler callout with the structure:

• text
• equation of different m_ij and m_ji
• plot 1
• text
• equation of equality m_ijN_i and m_jiN_j
• plot 2