Eurovision Analytics 🌐

This project sets up Neo4J and PostgreSQL databases containing participant and voting data from all Eurovision contests going back to 1956. You can use these and the provided services to answer interesting, complex questions about the contest over the years. See Use Cases for some cool examples.

Flyway is used for schema migrations and to import data from provided CSV files. These files are generated by the Eurovision Scraper companion project, which compiles Eurovision participant and voting data from Wikipedia.

The database is containerized using Docker Compose for easy setup and deployment.

Getting Started

1. Clone the repository:

   git clone https://github.com/jekrch/eurovision-analytics.git
   cd eurovision-analytics
   

2. Source data is included via the following CSV files, contained in the /data directory:

   • eurovision_vote_data.csv
   • eurovision_participant_data.csv

   You can either use the provided default source files or generate new data using my Eurovision Scraper project.

3. Start the containers:

   docker-compose up -d
   

   This command will start both databases, and run Flyway migrations, which generate the db schema and import data from CSV files. 3 interactive services will also be started which you can use to start exploring the data.

4. Connect to interactive data services

   Service	URL
   Frontend Home	http://localhost:3000
   GraphQL	http://localhost:4000
   Neo4J	http://localhost:7474

5. Connect to the PostgreSQL database:

   You can connect using a database client with the following details:

   • Host: localhost
   • Port: 5432
   • Database: eurovision
   • Username: postgres
   • Password: postgres

   The staging tables will be available in the eurovision database, and you can query them to access the imported data.

Neo4j graph data modeling 📊

Explore and analyze ESC contest data using graph-based queries and visualizations.

The Neo4j database is automatically set up and populated with data via docker-compose up -d. The provided docker-compose.yml file includes the necessary configuration to start a Neo4j container and execute the data migration scripts.

Access the Neo4j Browser:

1. Open a web browser and visit http://localhost:7474.
2. Click Connect. No credentials required

Graph-Based Analytics

With the Eurovision data loaded into Neo4j, you can perform various graph-based analyses and visualizations. Here are a few examples:

The top ten most prolific Eurovision songwriters

MATCH (sw:Songwriter)<-[:WRITTEN_BY]-(s:Song)
WITH sw, collect(s) as songs
ORDER BY size(songs) DESC
RETURN sw, songs
LIMIT 10

The top five winningest countries

MATCH (c:Country)<-[:REPRESENTS]-(s:Song)-[:PLACED]->(:FinalPlace {place: 1})
MATCH (s)-[:HAS_SONG]-(y:Year)
WITH c, collect({song: s, year: y}) AS songs, count(s) AS count
ORDER BY count DESC
LIMIT 5
RETURN c, songs

Countries with more than one song receiving Nil Poi in the finals

MATCH (c:Country)-[:REPRESENTS]-(s:Song)-[:RECEIVED]->(n:ZeroPoints), (s)-[:HAS_SONG]-(y:Year)
WITH c, s, y
ORDER BY c.name, y.year
WHERE y.year <> 1956 // points were not awared in 1956
WITH c, COLLECT(s) AS songs, COLLECT(y) AS years
WHERE SIZE(songs) > 1
RETURN c, songs, years

The top 3 running orders with the most songs winning the finals

MATCH (ro:FinalRunningOrder)<-[:RUNNING_ORDER]-(s:Song)-[:PLACED]->(fp:FinalPlace {place: 1})
MATCH (s)-[:PERFORMED_BY]->(a:Artist)
MATCH (s)-[:HAS_SONG]-(y:Year)
WITH ro, collect(s) as songs, collect(a) as artists, collect(y) as years
ORDER BY size(songs) DESC
LIMIT 3
UNWIND songs as s
UNWIND artists as a
UNWIND years as y
RETURN ro, s, a, y

PostgreSQL Use Cases 🔎

In addition to Neo4J, this project also provides a PostgreSQL database. You can answer all kinds of interesting, complex questions with normalized data. Here are several examples to get you started.

1. Which Eurovision songs were written in an imaginary language?

SELECT year, country, broadcaster, artist, song, language
FROM song_view
WHERE language = 'Imaginary'

year	country	broadcaster	artist	song	language
2003	Belgium	RTBF	Urban Trad	Sanomi	Imaginary
2008	Belgium	VRT	Ishtar	Julissi	Imaginary
...	...	...	...	...	...

2. Which artists competed in the most contests?

SELECT
  COUNT(*) AS song_count,
  artist,
  STRING_AGG(CAST(year AS TEXT), ', ' ORDER BY year ASC) as contest_years,
  STRING_AGG(DISTINCT country, ', ') as country
FROM song_view
GROUP BY artist
HAVING COUNT(*) > 1
ORDER BY COUNT(*) DESC;

song_count	artist	contest_years	country
4	Peter, Sue and Marc	1971, 1976, 1979, 1981	Switzerland
4	Valentina Monetta	2012, 2013, 2014, 2017	San Marino
4	Fud Leclerc	1956, 1958, 1960, 1962	Belgium
3	Hot Eyes	1984, 1985, 1988	Denmark
3	Lys Assia	1956, 1957, 1958	Switzerland
...	...	...	...

3. Which composers had the most songs in Eurovision?

SELECT
  COUNT(DISTINCT sv.song_id) AS song_count,
  c.name,
  STRING_AGG(CAST(sv.year AS TEXT), ', ' ORDER BY year ASC) as contest_years,
  STRING_AGG(DISTINCT sv.country, ', ') as country
FROM song_view sv
JOIN song_composer sc ON sc.song_id = sv.song_id
JOIN composer c ON c.id = sc.composer_id
GROUP BY c.name
HAVING COUNT(*) > 1
ORDER BY COUNT(*) DESC;

song_count	name	contest_years	country
24	Ralph Siegel	1974, 1976, 1979, 1980, 1980, 1981, 1982, 1985, 1987, 1988, 1990, 1992, 1994, 1997, 1999, 2002, 2003, 2006, 2009, 2012, 2013, 2014, 2015, 2017	Germany, Luxembourg, Montenegro, San Marino, Switzerland
16	Bernd Meinunger	1979, 1980, 1980, 1981, 1982, 1985, 1987, 1988, 1992, 1994, 1997, 1999, 2002, 2003, 2006, 2009	Germany, Luxembourg, Montenegro, Switzerland
12	Thomas G:son	2001, 2006, 2007, 2010, 2012, 2012, 2013, 2015, 2015, 2016, 2016, 2018	Cyprus, Denmark, Georgia, Malta, Spain, Sweden
9	Dimitris Kontopoulos	2009, 2013, 2014, 2016, 2017, 2018, 2019, 2021, 2021	Azerbaijan, Greece, Moldova, Russia
...	...	...	...

GraphQL API 🚀

This project also exposes the Neo4j Eurovision data through a GraphQL API, allowing you to query and retrieve data using GraphQL queries.

Accessing the GraphQL API

1. The GraphQL API is automatically set up and made available using Docker Compose. docker-compose up -d.
2. Use the API at http://localhost:4000

Example GraphQL Query

Here's an example GraphQL query that retrieves all songs from Croatia that have a final place:

query {
  songs(where: { country: { name: "Croatia" }, finalPlace: { place_NOT: null } }) {
    id
    name
    year {
      year
    }
    artist {
      name
    }
    finalPlace {
      place
    }
    totalPoints
  }
}