What do rugby fans actually think about rugby stadiums?

Experts at Liverugbytickets.co.uk have discovered the rugby stadiums around the world fans are most – and least – satisfied with. To do this, they created an index taking into consideration four factors: atmosphere, cleanliness, affordability and food. 

The study involved content analysis of over 15,000 TripAdvisor reviews for 40 of the biggest rugby stadiums by capacity. Stadiums were ranked in accordance with their average score out of ten across all four factors. 

Here are the best rugby stadiums… 

The rugby stadium fans are most satisfied with is Recreation Ground, situated in Bath, England. Home of Bath Rugby, the stadium scored a whopping score of 9.5 out of 10. The stadium had the highest score (9.7/10) for cleanliness and affordability. 

The stadium situated in second place is Welford Road, with a score of 9 out of 10. Located in Leicester, England, its highest rankings were for cleanliness (9.5/10) and food (9.3/10). 

In third place we have Totally Wicked Stadium, located in St. Helens, England, with a score of 8.5 out of 10. 

And the rugby stadium least liked by fans is… 

Estadio Ciudad de La Plata in Buenos Aires, Argentina is the rugby stadium fans are the least satisfied with. It scored only 0.3 points on average for the index. Estadio Ciudad de La Plata scored lowest for atmosphere, cleanliness and food (0.21 points each). 

Methodology  

1. Liverugbytickets.co.uk conducted research aimed at identifying the rugby stadiums fans are most – and least – satisfied with. The study involved content analysis of over 15,000 TripAdvisor reviews for 40 different stadiums. 

2. The sample of stadiums was compiled by consulting an official list of rugby union stadiums by capacity found on Wikipedia. 

3. Proceeding the collection of user reviews, pre-processing steps, including translation, text normalisation and tokenisation, were performed for more accurate results and faster computation. Text normalisation is a particularly important process, which re-casts words into their canonical form. 

4. The methods mentioned above are both standard natural language processing techniques employed to improve analytical accuracy. 

5. Text translation was performed with the use of Google Translate, a python library that uses Google Translate under the hood. The VADER sentiment analysis model was used to generate sentiment scores for each review. A customised lexicon was used to account for the semantics of the review sample to increase accuracy. 

6. While sentiment analysis models are vulnerable to inaccuracies, an inspection of scores given to random samples within the dataset were predominantly accurate. The method was therefore considered suitable for this application. 

7. To identify reviews with a specific context, a list of keywords synonymous with the following factors were used: atmosphere, cleanliness, affordability and food. These factors were arbitrarily chosen to reflect a good visitor experience. 

8. The average sentiment scores for reviews containing any of the keywords formerly identified within each context category were calculated. 

9. The resulting dataset contains the average sentiment score for each context category for each stadium. The final rankings were based on the average percentage rank for each category.