Visualization

Lattice provides two ways to visualize the resulting graph.

Python: Plotly figure

Lattice.visualize() returns an interactive Plotly figure. Use it in Jupyter notebooks or any Plotly-compatible environment.

fig = l.visualize()
fig.show()  # opens in browser

You can also load a saved lattice JSON instead of using the in-memory one:

fig = l.visualize(load_path="lattice.json")
fig.show()

The figure shows:

• Layer 0 (observations) — colored circle per observation
• Layer 1+ (insights) — one circle per insight per layer, with distinct colors
• Edges — gray lines connecting insights to their supporting nodes
• Hover text — full observation or insight text on hover

Pass the figure to Dash or Streamlit for embedding in a web app:

import dash
from dash import html, dcc

app = dash.Dash()
app.layout = html.Div([dcc.Graph(figure=l.visualize())])
app.run_server()

Browser: standalone HTML viewer

examples/visualize.html is a self-contained viewer that requires no server. Open it in any browser and upload a lattice.json file.

Features:

• Plotly.js canvas with scroll-to-zoom and drag-to-pan
• Three node types with distinct markers (observations, L1 insights, L2 insights)
• Click any node to open a detail panel on the right
• From an insight, see supporting observations or merged lower-level insights
• From an observation, see which insights reference it

To use:

1. Run l.save("lattice.json") after building
2. Open examples/visualize.html in your browser
3. Click Upload lattice.json and select your file