How We Analyze Channels
Every channel in the Project Profound archive receives a set of computed scores derived from our AI analysis pipeline. These scores help users understand a channel's strengths, focus areas, and how it compares to other channels in the archive. Here's exactly how each metric is calculated.
The Four Axes
Every channel is measured on four dimensions, grouped into two categories:
Research Elements
Intelligence Value
What it measures:How analytically deep and information-rich is this channel's content?
How we calculate it:Every video in our archive receives an Intelligence Value score (0–30) from our AI analysis pipeline. This score factors in claims density, specific programs and events mentioned, evidence quality, and overall analytical depth. A channel's Intelligence Value is the average of all its videos' scores, normalized to a 0–100 scale.
Speaker / Source Credibility
What it measures:How credible and well-sourced are the speakers and sources featured in this channel's content?
How we calculate it: A weighted composite of three signals:
- Source Diversity (40%)— How many unique persons of interest appear across the channel's videos. More diverse sourcing indicates broader investigative reach.
- Evidence Quality (40%) — Average evidence score from our encounter analysis. Channels featuring content with stronger verifiable evidence score higher.
- Program Depth (20%) — Average number of specific government programs mentioned per video. Channels discussing named programs (e.g., AAWSAP, AATIP, Project Blue Book) tend to have more substantive sourcing.
Encounter Elements
Encounter Depth
What it measures:How deep and detailed are the contact experiences featured in this channel's content?
How we calculate it:For every encounter we identify in a video, our AI pipeline scores it on a Contact Depth Scale (0–32) that evaluates four categories: observational detail, entity interaction, transcendent elements, and consciousness alteration. A channel's Encounter Depth is the average score across all its encounters, normalized to 0–100.
Impact
What it measures:How profoundly did the encounters featured in this channel transform the experiencers' lives?
How we calculate it:Our AI pipeline scores each encounter on a Transformation Scale (0–60) across multiple life domains: worldview, spirituality, relationships, career, psychological wellbeing, and more. A channel's Impact score is the average transformation score across all its encounters, normalized to 0–100.
Encounter Score vs Research Score
To make it easy to see a channel's focus at a glance, we combine the four axes into two composite scores:
Encounter Score
Average of Encounter Depth + Impact, then divided by the average channel's Encounter Score to produce a ratio. A score of 2.0× means this channel has twice the encounter content of the average channel. A score of 0.5× means half the average.
Research Score
Average of Intelligence Value + Speaker Credibility, then divided by the average channel's Research Score to produce a ratio. A score of 1.8× means this channel has nearly double the research depth of the average channel.
Why this matters:An encounter-heavy channel like "Experiencer Interviews" might score 3.2× on Encounter and 0.9× on Research — making it instantly obvious that this channel's strength is first-person accounts. A research channel like "National Geographic" might score 0.4× on Encounter and 1.3× on Research — clearly an analytical, investigation-focused channel.
Reading the Channel Focus Chart
The diamond-shaped chart on each channel page displays all four axes simultaneously. The chart is divided vertically:
- Left side = Encounter Elements (Impact + Depth). The shape extends further left on channels with strong encounter content.
- Right side = Research Elements (Intelligence + Credibility). The shape extends further right on channels with strong research content.
A balanced channel will have a symmetrical shape. An encounter-focused channel will lean heavily left. A research-focused channel will lean right. At a glance, you can see what kind of content a channel produces.
Archive Rankings
Every channel page includes a rankings box comparing it to all other channels in the archive. Here's what each metric means:
Archive Rank
Ranked by number of videos in our archive. Top 5 channels get a green badge, Top 10 get blue, Top 25 get bronze.
Views Rank
Ranked by total view count across all archived videos.
Engagement
Average comments-to-views ratio, shown as a multiple of the archive average. '2.3×' means this channel generates 2.3 times the comments relative to views compared to the typical channel.
Publishing Pace
How frequently the channel publishes, calculated as videos per month. Categorized as daily, weekly, bi-weekly, monthly, or sporadic.
Views per Video
Average views per archived video, shown alongside a comparison to the archive average.
The Channel Universe Map
The scatter plot on each channel page (and on the full universe map) plots every channel in the archive on two dimensions:
- X-axis: Speaker Credibility — how well-sourced is the channel
- Y-axis: Intelligence Value — how analytically deep is the content
- Dot size — reflects subscriber count (larger = more subscribers)
The map is divided into four quadrants by median lines: The Scholars (high intel, lower sourcing), The Authorities (high intel, high cred), The Explorers (narrative-focused), and The Broadcasters (wide sourcing, accessible format).
Guest Prominence Index (GPI)
The Guest Quality Over Time chart on each channel page tracks the caliber of guests and persons of interest featured on that channel year by year. It uses a composite metric called the Guest Prominence Index (GPI).
How GPI is Calculated
For each year, we identify every person of interest who appeared in the channel's videos that year. Each person contributes two signals:
- Credibility Score (60% weight) — Each person of interest in our archive has an
avg_credibility_score(0–85) computed from the evidence quality and sourcing standards of videos they appear in. This is normalized to a 0–100 scale. When credibility data is available, it receives 60% weight in the GPI calculation. - Cross-Archive Mentions (40% weight) — How many total videos across the entire archive mention this person (the
total_mentionsfield). This is normalized using a logarithmic scale to prevent outliers like frequently discussed figures from dominating. Higher mentions indicate a more prominent figure in the UAP discourse. When credibility data is available, this receives 40% weight.
Formula
normalized_cred = (avg_credibility_score / 85) × 100
normalized_mentions = (ln(avg_mentions + 1) / ln(max_mentions + 1)) × 100
GPI = normalized_cred × 0.6 + normalized_mentions × 0.4
(If no credibility data exists for that year's guests: GPI = normalized_mentions)
Rising GPI
The channel is increasingly featuring well-known, credible figures in the UAP field. This often correlates with the channel building legitimacy and industry connections over time.
Declining GPI
The channel may be shifting toward lesser-known guests or new voices. This isn't inherently negative — it could indicate the channel is platforming emerging experiencers not yet widely discussed elsewhere.
Data Sources & Transparency
All scores are computed by our AI analysis pipeline. Each video in the archive is analyzed for content type, entities mentioned, encounter details, evidence quality, and more. Channel-level scores are aggregated from individual video analyses.
Scores are refreshed regularly as new videos are added and analyzed. Because these are AI-generated assessments, they should be treated as analytical tools rather than definitive judgments. We continuously refine our analysis methodology.