Beach Dataset: Scientific Documentation

This document describes the structure and methodology of the Beach Dataset, which is part of the Beach Review Guide (BRG) Project, used in geographical, environmental, and tourism research. The dataset follows principles of transparency, verifiability, tiered confidence classification, and human-in-the-loop quality control.

Data Quality Assurance: Two-Stage Moderation

Each record is checked twice for consistency and accuracy:

1. Observer (Data Contributor) annotates fields strictly according to source evidence and project guidelines.
2. Moderator (Administrator) independently verifies every field against source material, checks geographic consistency using satellite imagery.

Only after successful moderation is a record assigned a Moderation Date and made available for public use or scientific analysis. Unmoderated or rejected records are excluded from all exports and visualisations.

Dataset Structure

The dataset contains two record types, collected using two related data collection procedures:

• Beach Profile — a static, long-term description of the beach’s physical and infrastructural characteristics. This profile is not tied to a specific visit and is intended to remain stable for several years.
• Visit Report — a time-specific observational record capturing environmental and social conditions during a particular visit.

These record types are stored separately in the database and serve different analytical purposes. Each field within either record type is assigned a confidence level (L1–L3) as defined below.

Data Collection Methodology

The methodology can be applied to any visual or narrative source that provides verifiable evidence of beach conditions. In the current implementation, the project relies mainly on publicly available user-generated videos from online platforms (primarily YouTube) due to their global coverage and accessibility.

Beach Profile

Beach Profiles are constructed using the following protocol:

• Analysis of ≥3 temporally distinct source videos to confirm stable characteristics.
• Verification of geographic coordinates and dimensions using satellite imagery (Google Maps, OpenStreetMap, SAT mode).
• Only characteristics consistently observable across videos are recorded. Transient features (weather, crowd) are excluded.

Visit Report

Visit Reports are derived from individual source videos that satisfy the following inclusion criteria:

• Depicts the beach surface and nearshore zone from ground-level perspective.
• Contains extractable visit date (year, and month if explicitly stated).
• Is not a Short or promotional montage.
• Provides audible or visual evidence for claimed conditions (e.g., wind, noise, crowd).

Each Visit Report is linked to one Beach Profile via beach_id.

Annotation Protocol

The detailed visual criteria used to distinguish categorical values are defined in an internal annotation guide available only to project contributors (observers). External users can treat categorical fields as discrete observational classes without needing access to the internal guidelines.

Limitations

The Beach Dataset, while methodologically rigorous within its design constraints, is subject to several limitations that affect its scope, representativeness, and interpretability:

• Non-representative sampling: the dataset reflects only beaches that appear in publicly available videos. Remote or less frequently filmed locations may be underrepresented, which affects geographic coverage.
• Temporal non-uniformity: Visit Reports are distributed irregularly across months and years, with clustering during peak tourist seasons. This limits the ability to perform consistent seasonal or interannual trend analysis.
• Observational constraints: All L1 fields rely on passive visual/auditory evidence from uncontrolled user-generated content. Key limitations include:
  • inability to confirm absence (e.g., “no toilets” requires full-beach visibility, which is rare);
  • limited depth of field (water bottom, wave height often inferred from surface view only);
  • no ground-truthed measurements (e.g., temperature, wave height are estimated, not instrumentally recorded).
• Observer subjectivity: Despite standardized protocols and two-stage moderation, categorical judgments involve perceptual interpretation, introducing inter-observer variance.
• Platform fragility: Source videos may be deleted, privatized, or altered by uploaders, affecting long-term auditability of individual records.

The dataset is most suitable for descriptive comparisons, cross-site characterization, and preliminary hypothesis generation within the spatio-temporal limits of the source material.

Data Quality Framework

Within each record type, every field is assigned one of three confidence levels, informed by common practices in geographic data quality assessment:

Field Definitions and Classification

Table 2 lists fields, their record type affiliation, scientific definition, permissible values, source, and confidence level.

Data Availability, Use, and Attribution

The L1 and L2 portions of the Beach Dataset are not publicly downloadable at this time. They may be made available under a CC BY 4.0 license for non-commercial academic research upon request. The L3 layer (editorial content) is excluded from redistribution and is intended solely for user-facing presentation on the Beach Review Guide (BRG) platform.

Commercial use of any part of the dataset—including licensing, embedding in websites or mobile applications, use in commercial products, or resale—requires a separate written agreement with the project author.

To request access to the dataset or inquire about commercial licensing and collaboration, please contact: coconut@beachreviewguide.com.

For academic citation, please use the following format:

Vityasev, Y. M. (2026). Beach Dataset. Beach Review Guide. https://beachreviewguide.com/dataset-description/

Methodology Attribution and Intellectual Property

The data collection procedures, field definitions, confidence system (L1–L3), and moderation workflow constitute the methodological basis of the BRG project. The custom categorical scales (e.g., wave height, wind intensity, infrastructure, cleanliness) were designed specifically for visual assessment from user-generated videos. The methodology is an original work; using it outside the BRG project requires permission from the author.

Proper attribution is required in any derivative academic work that references this methodology:

“Data collection followed the Beach Review Guide (BRG) methodology (Vityasev, 2026).”