Source-linked AI summary
Exploring ESC Winners with Nested Diagrams
Anurag Sharma, Marcel Nöhre, Gerd Stumme
TL;DR
Understanding how ESC winners’ voting support relates to their musical characteristics requires a framework that represents both dimensions together. ConceptFlow constructs nested line diagrams that reveal dataset-level implications between voting patterns and musical properties.
Problem
The paper examines relationships between ESC winners’ voting patterns and musical characteristics within a unified conceptual structure.
Method
ConceptFlow applies conceptual scaling to many-valued ESC winner contexts, computes separate factor lattices, and combines them in interactive nested line diagrams.
Results
The diagram reveals implications linking voting support with tempo and key, including dependencies involving cultural, regional, political, and historical support.
Takeaways & Limitations
Nested line diagrams make relationships and implications across multiple conceptual scales directly accessible in the ESC case study.
Takeaways & Limitations
The ESC analysis demonstrates ConceptFlow’s capabilities rather than providing a comprehensive statistical study of voting behaviour, and voting profiles combine data differently before and after 2016.
Abstract
from arXiv · showhide
We present ConceptFlow, a scikit-learn-compatible Python library for Formal Concept Analysis that constructs and renders nested line diagrams from many-valued formal contexts. Given a many-valued context and a partition of its attributes into conceptual scales, ConceptFlow performs conceptual scaling, computes the factor lattices, identifies filled nodes of the corresponding subdirect product, and produces an interactive visualization. We apply ConceptFlow to the winners of the Eurovision Song Contest from 1975 to 2025, exploring relationships between voting patterns and musical characteristics. Voting support is captured by an outer scale spanning regional, cultural, historical, and political dimensions, while an inner scale captures musical characteristics via tempo and key. The resulting nested line diagram reveals implications across both scales, exposing dependencies between how winning entries were voted for and the musical properties they share.
1 Introduction
This paper presents ConceptFlow’s nested line diagram functionality for analyzing Eurovision winners through Formal Concept Analysis. It combines voting-support patterns and musical characteristics to reveal implications between how winning entries were voted for and the musical properties they share.
- 1 Introduction: Since 1975, ESC countries have distributed scores of 1-8, 10, and 12 to their ten favourite performances, without voting for themselves.Score-deciding mechanisms have changed repeatedly among juries, public televoting, or combinations of both.
- 1 Introduction: ConceptFlow is a scikit-learn-compatible Python library for Formal Concept Analysis with nested line diagram functionality.Its components follow the scikit-learn estimator interface.
- 1 Introduction: The outer scale captures voting support across taxonomic clusters, while the inner scale captures musical characteristics.
- 1 Introduction: The resulting formal context reveals implications between voting patterns and the musical properties shared by winning entries.
2 Background
Formal Concept Analysis represents concepts as a lattice ordered by extent inclusion, while nested line diagrams decompose attributes into scales and visualize the resulting factor-lattice structure.
- Formal Concept Analysis: A formal context comprises objects, attributes, and an incidence relation; formal concepts pair maximal object sets with their shared attributes and form a concept lattice.Concepts are ordered by extent inclusion.
- Factorization: Apposing outer and inner contexts yields a concept lattice isomorphic to a subdirect join-semilattice product of their factor lattices.Each combined formal concept corresponds uniquely to one outer concept and one inner concept.
- Line Diagrams: Line diagrams represent concepts as nodes and connect pairs in the transitive reduction of the lattice order, placing smaller concepts below larger ones.Node coordinates in R2 determine positions in the drawing.
- Nested Line Diagrams: Nested line diagrams decompose attributes into scales, display one factor lattice outside, and place a copy of another factor lattice inside each outer concept.The subcontexts share the same object set.
3 ESC Dataset and Scale Construction
The study analyzes 50 Eurovision winners from 1975–2025, excluding 2020, using voting profiles and musical metadata to construct a many-valued context. Conceptual scaling represents voting affinities through an outer scale and tempo/key through an inner scale.
- Dataset: 50 Eurovision winners from 1975–2025 are analyzed, excluding the cancelled 2020 contest, using publicly available voting results and song metadata.Metadata includes each song’s BPM and key.
- Dataset: A single voting profile uses the combined score before 2016 and, thereafter, whichever of jury or televote awarded the larger total number of points.Before 2016, the contest reported only combined voting; since 2016, jury and public televoting are separate.
- Scale construction: The resulting many-valued context is the input to the conceptual scaling process.The context combines voting information with musical properties.
- Outer scale: The outer scale encodes voting support from four manually clustered tie types: regional, cultural, historical, and political.The country clusters reflect commonly discussed sources of voting affinity.
- Inner scale: Tempo uses an ordinal threshold scale at 100 and 150 BPM, while key uses mutually exclusive major and minor attributes.Songs at least 150 BPM also satisfy the ≥100 BPM attribute.
4 Constructing the Nested Line Diagram
ConceptFlow constructs nested line diagrams by scaling a many-valued context into outer and inner formal contexts, then computing their factor lattices. It identifies filled nodes through coordinate-pair joins and renders the resulting fixed-layout structure interactively.
- Conceptual scaling: ConceptFlow is a scikit-learn-compatible Python library providing Formal Concept Analysis basics, conceptual scaling, and visualization tools.For a many-valued formal context, it independently constructs outer and inner formal contexts and computes their corresponding concept lattices.
- Conceptual scaling: The 50 ESC winners are scaled into Kouter and Kinner, with outer attributes containing country clusters and inner attributes {≥100 BPM, ≥150 BPM, minor, major}.The outer attributes comprise all country clusters from the four clusterings.
- Factor lattices: The concept lattices of Kouter and Kinner serve as the nested diagram’s factor lattices because both contexts share the same object set and satisfy the apposition condition.The shared object set is G, consisting of the ESC winners.
- Filled-node construction: Filled nodes are computed directly by generating object-concept coordinate pairs, adding (⊥outer, ⊥inner), and repeatedly taking componentwise joins to a fixpoint.The explicitly added bottom pair represents (M ′, M), the join of the empty set, and the fixpoint equals the subdirect join-preserving embedding’s image.
- Visualization: DimFlux lays out each factor lattice once, reusing the inner layout as a fixed template so outer differences appear only through filled nodes.The completed structure is exported as JSON and rendered through an interactive D3.js visualization independent of the FCA construction algorithms.
5 Results
ConceptFlow’s nested line diagram reveals dependencies between voting support and musical characteristics among 50 Eurovision winners. It localises these implications, showing where musical combinations require additional regional, cultural, political, or historical support.
- Voting and musical dependencies: Among 50 ESC winners, every song with a tempo of at least 150 BPM also received cultural support.This dependency is visible as an implication in the nested line diagram.
- Voting and musical dependencies: Among winners receiving regional or cultural support, every minor-key song has a tempo of at least 100 BPM.The corresponding absent node represents this dependency between tempo and key mode.
- Combined voting-support relationships: Winners combining cultural support, a minor key, and a tempo of at least 150 BPM also receive political and historical support.The absent node indicates that this combination does not occur without those additional voting-support attributes.
- Localisation of implications: The nested representation localises implications within particular subdirect-product nodes, showing where musical combinations are possible or excluded by additional voting-support conditions.This visual localisation replaces an exclusively algebraic listing of implications.
6 Conclusion
ConceptFlow provides a scikit-learn-compatible library for constructing interactive nested line diagrams that connect voting patterns and musical characteristics across conceptual scales. Future work will add scales, implication theory, and richer interactive visualization.
- Contributions: ConceptFlow supports conceptual scaling, concept lattice construction, and interactive nested line diagrams for Formal Concept Analysis.The library is scikit-learn-compatible and combines multiple conceptual scales in a single visualization.
- Case study: The Eurovision case study makes relationships and implications between voting patterns and musical characteristics directly accessible through nested line diagrams.The diagrams combine multiple conceptual scales into one visualization.
- Future work: Future work will add conceptual scales, implication theory, and further interactive visualization capabilities.Planned developments include canonical implication bases and interface support for inspecting implications associated with concepts and unfilled diagram regions.