Chord Diagram

# === IMPORTS ===
import pandas as pd
import holoviews as hv
from holoviews import opts, dim
from bokeh.plotting import show
hv.extension('bokeh')

# === USER-EDITABLE PARAMETERS ===
# Change: List of region names (length determines number of nodes)
regions = [
    'North America',
    'Europe',
    'Latin America',
    'Asia',
    'Africa',
    'Middle East',
    'Oceania',
    'South Asia'
]

# Change: List of raw migration flows as (source_index, target_index, value_in_millions) tuples
# Indices must be integers from 0 to len(regions)-1
links_raw = [
    (0, 1, 0.5),   # North America → Europe
    (1, 0, 0.8),   # Europe → North America
    (2, 0, 1.5),   # Latin America → North America
    (3, 0, 1.2),   # Asia → North America
    (4, 1, 0.9),   # Africa → Europe
    (3, 1, 0.7),   # Asia → Europe
    (5, 1, 0.4),   # Middle East → Europe
    (6, 3, 0.3),   # Oceania → Asia
    (1, 4, 0.6),   # Europe → Africa (additional)
    (1, 2, 0.2),   # Europe → Latin America (additional)
    (0, 3, 0.4),   # North America → Asia (additional)
    (7, 1, 0.5),   # South Asia → Europe (additional)
    (2, 3, 0.4),   # Latin America → Asia (additional)
    (4, 5, 0.2),   # Africa → Middle East (additional)
    (5, 4, 0.3),   # Middle East → Africa (additional)
    (6, 0, 0.1),   # Oceania → North America (additional)
    (3, 6, 0.2),   # Asia → Oceania (additional)
    (7, 0, 0.6)    # South Asia → North America (additional)
]

title = "Global Migration Patterns (Annual Flows in Millions)"  # Change: Plot title
chord_width = 1000  # Change: Width of the chord diagram (increased for better visibility)
chord_height = 1000  # Change: Height of the chord diagram (increased for better visibility)
cmap = 'Category20'  # Change: Colormap for categorical colors (regions); use 'Category20', 'Set3', etc.

# === Data Preparation ===
# Compute total volumes (in + out) for ordering regions by migration volume
totals = {i: {'out': 0.0, 'in': 0.0} for i in range(len(regions))}
for source, target, value in links_raw:
    totals[source]['out'] += value
    totals[target]['in'] += value
total_volumes = {i: totals[i]['out'] + totals[i]['in'] for i in range(len(regions))}

# Order regions by total volume descending
sorted_indices = sorted(range(len(regions)), key=lambda i: total_volumes[i], reverse=True)
index_mapping = {old_idx: new_idx for new_idx, old_idx in enumerate(sorted_indices)}

# Create nodes DataFrame (MUST have 'index' column with sequential integers 0-N)
nodes_df = pd.DataFrame({
    'index': range(len(regions)),
    'name': [regions[old_idx] for old_idx in sorted_indices]
})

# Remap links to new sorted indices
links_list = []
for source, target, value in links_raw:
    new_source = index_mapping[source]
    new_target = index_mapping[target]
    links_list.append({
        'source': new_source,
        'target': new_target,
        'value': value
    })
links_df = pd.DataFrame(links_list)

# === Print Relevant Statistics ===
print("=== Migration Analysis (Annual Flows in Millions) ===")
print("Total volumes by region (in + out, sorted descending):")
for new_idx in range(len(regions)):
    old_idx = sorted_indices[new_idx]
    volume = total_volumes[old_idx]
    region_name = nodes_df.loc[new_idx, 'name']
    print(f"{region_name}: {volume:.1f}M (out: {totals[old_idx]['out']:.1f}M, in: {totals[old_idx]['in']:.1f}M)")

print("\nAll migration flows:")
for _, row in links_df.iterrows():
    source_name = nodes_df.loc[row['source'], 'name']
    target_name = nodes_df.loc[row['target'], 'name']
    print(f"{source_name} → {target_name}: {row['value']:.1f}M")

print(f"\nTotal global migration volume: {sum(total_volumes.values()):.1f}M")
print(f"Highest volume region: {nodes_df.loc[0, 'name']} ({max(total_volumes.values()):.1f}M)")

# === Create Chord Diagram ===
# CRITICAL: nodes must use hv.Dataset with 'index' as key dimension
nodes = hv.Dataset(nodes_df, 'index')

# Create Chord: (links_df, nodes), kdims=['source','target'], vdims='value'
chord_plot = hv.Chord((links_df, nodes), ['source', 'target'], 'value')

# Style: color arcs by origin (source), nodes by index; proportional widths; hover tooltips; node labels
# Improvements: larger fonts, more padding for labels, explicit label styling, interactive tools, larger nodes/chords
styled_chord = chord_plot.opts(
    opts.Chord(
        title=title,
        cmap=cmap,
        edge_cmap=cmap,
        edge_color=dim('source').astype(str),  # Color arcs by origin region
        node_color=dim('index').astype(str),   # Color nodes by region
        labels='name',                         # Label nodes with region names
        padding=0.35,                          # Increased padding for label space
        node_size=20,                          # Larger nodes for better visibility
        edge_line_width=4,                     # Thicker chords for emphasis
        width=chord_width,
        height=chord_height,
        fontsize={
            'title': 22,     # Larger title font size
            'labels': 18,    # Larger node labels font size
            'ticks': 14      # Tick labels (if any)
        },
        # Enhanced hover and interaction tools
        tools=['hover', 'pan', 'wheel_zoom', 'reset']
    )
)

renderer = hv.renderer('bokeh')
fig = renderer.get_plot(styled_chord).state
show(fig)
plt.show()
# END-OF-CODE

Total Annual Migration: 18.9M people
Top Source Region: Asia
Top Destination: North America