How We Build
Our Data

Accurate dive information. Verified by people and powered by technology.

Scattered, outdated, and conflicting information can make dive planning risky. We combine human expertise with advanced technology to deliver the most reliable and up-to-date dive site data in the world.

Home/About/How We Build Our Data

The Problem with Dive Data

Dive site information is scattered across forums, blogs, outdated directories and maps. Names are inconsistent, coordinates are wrong, and critical details like depth or hazards often conflict. In diving, bad data isn't just inconvenient—it can be dangerous.

Inconsistent site names and locations
Conflicting depths, conditions and hazards
Outdated or unverified information
No clear way to know what to trust

1. Our Data Sources

We don't rely on a single source. We aggregate from multiple trusted inputs.

Dive Operator Submissions

Local dive centers and operators provide first-hand knowledge about the sites they dive every day. This source has the highest trust in our system.

Highest Trust

Structured Datasets

We use open geospatial data, government records and authoritative marine datasets as our foundation for accurate mapping and baseline information.

High Trust

Community & AI Aggregation

We crawl forums, blogs, reviews and directories, then use AI to extract useful facts and turn unstructured text into structured data.

Medium Trust

2. Our Data Pipeline

Every piece of data goes through a rigorous process before it appears on our site.

1. Ingest

Collect raw data from all sources.

2. Normalize

Standardize names, formats and coordinates.

3. Deduplicate

Detect and merge duplicate sites using geospatial clustering (PostGIS).

4. Extract Facts

AI models extract depths, conditions, marine life, hazards and more.

5. Score & Verify

Facts are scored, verified and published based on trust level.

3. Our Verification System

Not all data is treated equally. Our Trust Score tells you how much confidence you can have.

Level 3
Verified

(High Trust)

Confirmed by a verified dive center or supported by strong visual evidence or official data.

You can trust this.

Level 2
Consensus

(Med Trust)

Supported by multiple independent sources that strongly agree.

Generally reliable.

Level 1
Unverified

(Low Trust)

Based on a single uncorroborated report or weak evidence.

Use with caution.

Level 0
Rejected

Information flagged by the community as dangerous, false or inherently misleading.

We don't publish this.

4. Community Flags & Reviews

Divers help keep our data honest. If something looks wrong, you can flag it on any dive site page.

1. You flag an issue

Report incorrect info, outdated conditions or safety concerns.

2. We freeze the fact

The related fact's Trust Score is temporarily frozen.

3. In review queue

Our team reviews the report and gathers more evidence.

4. Decision made

Fact is updated, verified or rejected. The score is adjusted.

Your reports and local knowledge directly improve the safety and accuracy of the entire community.

How We Build Our Data