Network Community Detection · Eurovision · Data Analysis

Sixty Years of Eurovision Voting Blocs

How national voting allegiances cluster, fracture and re-form across the contest, from 1957 to 2022

Alphab.: Hagendorfer Simon, Kaufmann Thea & Wegendt Ellen · {e12008552, e12002221, e12530032}@student.tuwien.ac.at

Abstract

Each contest is treated as a directed, weighted network in which juries and televoters award points to one another, and we analyzed the voting communities in decade intervals. Tracing these communities forward reveals coherent blocs that persist for generations — a stable Nordic–Atlantic core, a Mediterranean cluster, and an Eastern bloc that consolidates sharply after 2000.

online visualization and resources

The evolution of voting communities

each thread = one country · colour = bloc lineage · height ∝ bloc size

Cross-bloc voting in the recent contests

column = country · background = community that year · bar = jury+public votes by recipient community

The diagram shows how countries allocate votes between communities. The wider the portion of the bar, the more outgoing votes this country gave to the community of this color. Countries may still be assigned to a community if they receive substantial support from its members or if their voting behavior does not consistently fit a single group. Communities were determined on multiple year aggregated voting data (to match the visualization above), reflecting long-term trends rather than year-to-year variations.

The Blocs — 2022

Western & Mediterranean

GE AL LU FR NL BE AM CY GR IL ME DE
Southern & Eastern

AZ MD BG SM RO ES IT PT CH MK
Nordic & Central

MT RS AT CZ SI HR NO FI AU EE SE IS DK
Atlantic & Eastern frontier

UA PL LT IE GB LV

Colours trace each bloc’s lineage backward through the decades by membership overlap.

How to read it

Panel I — every country is a thread. Where threads bundle into a coloured block, those countries formed one community for that decade. Follow a colour left→right to watch a bloc survive, absorb or shed members.

Panel II — each column is a country; background colour shows its community for that contest year. The stacked bar in each cell shows the distribution of jury and public points it awarded, broken down by recipient community, 2021–2025.

Key findings

A Nordic–Baltic community is the most durable structure in the contest, recognisable in every snapshot from the 1980s onward.
An Eastern–European bloc crystallises after 2000 as the contest expands, growing to fifteen members by 2017.

Method

There are two phases: We first found those pairs of countries that voted for each other more than we would expect (i.e. some behaviour that can't be explained by chance) using a simulation. Then we ran a community detection algorithm on the resulting graph, which finds out where there is a lot of collusion (voting for each other) in comparison to all the voting behaviour in the graph.

Abstract

I The evolution of voting communities

II Cross-bloc voting in the recent contests

The evolution of voting communities

Cross-bloc voting in the recent contests