What This Chart Shows
This chart shows your causal influence network as an animated “brain”: nodes are grouped by channel into spherical lobes on a larger sphere, and edges show causal flow between those channels. Over time, edges and node activity (glow and size) change by window so you can see when influence is strong or weak and how connections form or fade. This chart allows you to see how channels influence each other over time and which lobes are most active in each period.
Key Questions This Chart Helps Answer
- How do channels (lobes) influence each other over time?
- When is a channel “hot” (high activity) vs “cold” in the causal network?
- Which causal links between channels appear, persist, or fade by window?
- How is the causal graph structured by channel rather than by tier?
Axes, Metrics, and Units
Element | Description |
|---|---|
| Lobes (clusters) | Spherical clusters of nodes; one lobe per channel. Lobe centers are on a large sphere (brain shell); nodes sit on a smaller sphere inside each lobe. |
| Node position | Fixed per node across frames (lobe layout is stable). |
| Node size (3 layers) | Outer glow, inner glow, core. Sizes scale with activity in that window. Larger = more incident edge weight in that window. |
| Node color | By channel (lobe). Alpha of outer/inner layers scales with activity (brighter = more active). |
| Edges (3 layers) | “Synaptic” look: wide faint cyan, mid blue, thin bright core. Vary by frame: only edges valid in the TCEG window are shown. Up to 200 edges per window. |
| Frame | One time window. Date (or range) shown in slider as “Signal: …”. |
| Playback | Fire = play, Hold = pause. Speed: Think (slow), Active, Rapid, Hyper (fast). Slider = jump to a window by date. |
Positions are layout coordinates (channel-based spheres). Edge presence is model-derived from TCEG validity; node activity is derived from edge weights in that window.
Control Options Reference
Control | Meaning |
|---|---|
| Fire | Start animation; frames advance automatically. |
| Hold | Pause at the current frame. |
| Think / Active / Rapid / Hyper | Animation speed (Think = slowest, Hyper = fastest). |
| Slider (Signal) | Scrub to a specific time window; label shows the window’s date or range. |
How to Interpret the Results
- Lobes = channels: Each glowing cluster is one channel (or channel group). Proximity is layout, not necessarily “closeness” in the funnel.
- Brightness and size: Brighter or larger nodes = higher activity in that window (more causal influence). Dim or small = less.
- Edges between lobes: Lines between clusters show cross-channel causal links. Edges appearing or disappearing = relationships turning on or off in that window.
- Animation: Play to see how activity and connections change over time; use the slider to align with campaigns or external events.
- Stable vs transient: Lobes and links that stay bright and connected across many windows = stable structure; ones that come and go = transient.
Practical Applications for Marketers
Application | How to use this chart |
|---|---|
| Channel interplay | See which channels influence which others and how that changes over time. |
| Timing and campaigns | Align “hot” lobes and new edges with campaign start/end or seasonality. |
| Structural shifts | Spot windows where many links or lobes change (e.g., data or market breaks). |
| Storytelling | Use the “brain” metaphor to explain that attribution is a live, connected system. |
| Planning | Use stable vs transient patterns to prioritize always-on vs time-bound tactics. |
Common Mistakes & Misinterpretations
Mistake | Why it is a problem | How to avoid |
|---|---|---|
| Taking “lobe” position as importance | Position is layout (sphere placement), not performance or priority. | Use activity (brightness/size) and edges to judge influence. |
| Treating “edge appears” as proven new cause | Validity is from the TCEG model; it’s when the link is valid in the model, not necessarily when the true cause started. | Use as a signal; validate with tests or other analyses. |
| Assuming channel = business channel | “Channel” here is a grouping from the graph (e.g., degree-based or kernel-defined); it may not match your org’s channel list 1:1. | Check how channels/lobes are defined in your setup. |
| Over-reading one frame | One window can be noisy or atypical. | Watch several frames and use the slider to compare windows. |
| Ignoring window length | Each frame is an aggregate over a time window; short-lived effects may be smoothed. | Be aware of window size when interpreting “when” something changes. |
Caveats & Considerations
- Minimum windows: Needs at least 2 time windows for animation. With only one, the chart shows “Need at least 2 windows for animation (have W).”
- Edge limit: At most 200 edges per window are shown. Dense graphs are simplified per frame.
- Channel definition: Lobe/channel grouping can come from the world-model kernel or fallbacks (e.g., degree-based quantiles or even splits). It may not match your business channel taxonomy exactly.
- Metaphor vs analytics: The “neural” look is a visualization metaphor. Underlying logic is the same causal graph and TCEG validity as other network views; use it for intuition and narrative, not as a different model.
- Performance: 3D animation with glow and many nodes/edges can be heavy on some devices or browsers.
- Accessibility: Dark theme and reliance on color/glow may be hard for some users; pair with other charts or tables for key numbers.