Dataset Integration & Bottom-Up Ontology – Level II

After creating our first ontology schema, we moved to a bottom-up refinement using real-world data. The goal of this phase was to validate, expand, and ground the ontology with empirical information from actual portrayals of LGBTQ+ characters.

🗃️ Dataset Composition

Since no existing dataset met our requirements, we created a custom dataset featuring 29 LGBTQ+ characters from teen dramas and adjacent series.

Each entry captures detailed attributes about the character, their storyline, the show context, and the representational quality.

Key dimensions included:

  • Representation Type (Fair / Unfair)

  • Character depth, role, and resolution

  • Identity markers: gender, orientation, transition

  • Portrayal structure: authenticity, stereotypes, normalization

  • Creator, performer, budget, release year, production context

This dataset became the empirical basis for the second level of modeling.

🧩 From Data to Ontology Enrichment

The dataset allowed us to:

  • ✅ Validate the structure of the original ontology

  • ✅ Add new classes and subclasses based on repeated patterns

  • ✅ Refine object and data properties with more nuance

  • ✅ Introduce scores, roles, and context-based subclasses

📷 Ontology Visualization

Bottom-Up Visualization (📁 download the xmind file)

🆕 New & Enriched Classes

This is a list of the classes and relative sublasses.

TVshow

  • Country Specifies the country where the TV show was produced or originally aired. This is relevant for understanding cultural and regulatory contexts that influence LGBTQ+ representation.

  • Genre Denotes the genre of the TV show (e.g., drama, comedy, fantasy), which can affect the tone, tropes, and thematic treatment of LGBTQ+ characters.

Person

  • Gender Represents the gender identity of a person (creator or performer). Tracking gender helps explore who gets to tell queer stories and how gender dynamics shape representation.

  • SexOrientation Describes the sexual orientation of the creator or performer, where disclosed or relevant. This class allows analysis of whether LGBTQ+ roles are portrayed or written by members of the community.

  • CharacterDepth Assesses how psychologically and narratively developed a character is. This includes dimensions such as backstory, emotional complexity, and agency in the plot.

PlotResolution

  • NormalizationOfRelationship Indicates that an LGBTQ+ relationship is treated as natural and unproblematic within the narrative. This contributes to positive cultural messaging and reduces stigma.

  • AffirmationOfIdentity Signifies a resolution in which the character’s queer identity is accepted, celebrated, or integrated into the storyline in a meaningful way.

NarrativeRole

  • CharacterTransformation Highlights whether the character undergoes significant development, self-discovery, or change in values or relationships — particularly in relation to their identity.

  • Dynamic Labels characters whose narrative arc shows growth, change, or evolving relationships, particularly those tied to identity or conflict resolution.

  • Static Marks characters who remain unchanged throughout the story, often used to explore token roles or those lacking emotional depth.

Stereotype

  • GayBestFriend A common trope where the LGBTQ+ character serves as a sidekick to the heterosexual protagonist, often without a fully developed personal storyline or emotional life.

  • TragicTrope Refers to stereotypical narratives in which LGBTQ+ characters face suffering, trauma, or death — often reinforcing negative associations or victimhood.

🔗 New Object Properties

Property
Domain → Range
Description

hasGender

Character → Gender Person →Gender

Tracks gender identityAssigns a gender identity to either a real person (creator or performer) or a fictional character. This enables analysis of gender representation both on and off-screen.

hasSexOrientation

Character → SexOrientation Person → SexOrientation

Describes the sexual orientation of creators, performers, or characters. It supports the study of authentic casting and queer identity portrayal in media.

hasCharacterDepth

Character → CharacterDepth

Indicates the level of narrative and psychological development a character has. It helps distinguish between flat and multidimensional portrayals.

schema:countryOfOrigin

TVShow →Country

Specifies the country where the TV show was originally produced or broadcast. This helps contextualize cultural norms and production standards influencing LGBTQ+ portrayals.

associatedWith

UnfairRepresentation → NegativeRoleModel

FairRepresentation → PositiveRoleModel

Links a type of representation to the kind of role model it produces. Fair representations are associated with positive role models, while unfair ones are linked to negative role models.impactCommunity

impactCommunity

PortrayalofLgbtqCharacter → LGBTQCommunityImpact

Connects a portrayal of an LGBTQ+ character to its broader social impact. It reflects how the portrayal influences the LGBTQ+ community and public perception.

📊 New & Enriched Data Properties

Property
Domain
Range (Examples)

schema:datePublished

TVshow → Date

Records the original release date of the TV show. This is useful for analyzing trends in LGBTQ+ representation over time.

schema:genre

TVshow → Genre

Specifies the genre category of the show (e.g., drama, comedy, sci-fi). Genre plays a role in shaping tone, stereotypes, and the depth of representation.

hasTitle

TVShow → Title

Captures the official name of the TV show. This property provides the basic identifying label for any media entity in the ontology.

isTransgender

Gender → Boolean

A boolean flag indicating whether a gender identity is transgender. This allows for the distinction of transgender representation within the broader gender category.

📄 New Instances

As we moved into the second level of modeling, we expanded our ontology with more refined instances to support deeper analysis. These additions are crucial for distinguishing between different forms of representation and their narrative depth.

  • LGBTQPortrayal:detailed (rdf:type) Indicates a character with narrative complexity, emotional nuance, and evolving motivations. These characters are typically involved in multiple plot layers and contribute meaningfully to the story arc.

  • LGBTQPortrayal:superficial (rdf:type) Describes a character with minimal development, often existing to fulfill a trope or secondary role. These portrayals lack depth and are frequently one-dimensional or tokenistic.

These instance-level refinements allow us to go beyond binary categories like fair/unfair and explore the quality, depth, and function of LGBTQ+ portrayals within each series.

🧠 Example: Enrichment from Dataset

Character: Jules Vaughn (Euphoria) RepresentationType: Fair CharacterDepth: Detailed AffirmationOfIdentity: Yes NarrativeRole: Dynamic Stereotype: Avoided PlotResolution: Open-ended (personal growth) Impact: Positive Role Model isTransgender: Yes Portrayal Lens: Social Impact Creator Gender: Male Platform: HBO

This entry provided direct grounding for multiple ontology elements, such as characterTransformation, affirmationOfIdentity, and normalizationOfRelationship.

💬 Why Bottom-Up Matters

This step ensured that our ontology wasn’t only theory-driven, but also responsive to real-world patterns. It allowed us to represent characters as they appear — not just as ideal models — and to surface contradictions, gaps, or inconsistencies in mainstream media.

It also enabled quantitative queries over the ontology and dataset, such as:

  • 🔍 How many trans characters are portrayed fairly?

  • 🔍 Is there a correlation between creator identity and authenticity level?

  • 🔍 Which portrayals include stereotype tropes and still resolve positively?

This bottom-up modeling ensures the ontology remains grounded, scalable, and usable across diverse cultural and media studies use cases.

PreviousConceptual Modeling – Level INextDataset Schema & Mapping Notes

Last updated