By Caio Mello, Doctoral Researcher at the School of Advanced Study, University of London

This blogpost is a summary of the presentation I delivered with my colleague Daniela Major in the conference Engaging with Web Archives: ‘Opportunities, Challenges and Potentialities’ in September 2020. This presentation is entitled ‘Tracking and analysing media events through web archives’.

My research explores the media coverage of the Olympic Games in a cross-cultural, cross-lingual and temporal perspective. I am especially interested in comparing how the concept of 'Olympic legacy' has been approached by the Brazilian and British media considering different locations, languages and social-political contexts. I have written a bit about this before on the UK Web Archive blog in December 2019 and March 2020.

Because of its controversial nature, the term Olympic legacy is used in a variety of contexts and it has multiple meanings. Considering its narrative importance to legitimize the billionaire investment of cities to host these events, this study has as the main objective to explore and define the concept of Olympic Legacy and how it changes over time.

Here however, I will be focusing on my experience doing a secondment at the British Library with the UK Web Archive team. I have explored the potential of using the platform Shine to track news articles on Olympic legacy.

Why Shine?

Shine is a tool to explore .uk websites archived by the Internet Archive between 1996 and April 2013. While a big part of the content of the UK Web Archive can only be accessed from inside the British Library, Shine is open access and provides us with search results and URL data that can be easier to manage.

We have developed a pipeline based on 5 steps: searching, extraction, cleaning, filtering and visualisation. To extract information, we have conducted web scraping of the data using Python notebooks looking at specific newspapers (like The Guardian) and broadcast websites (like BBC) using the keyword “Olympic legacy”. Having searched for URL’s in Shine and extracted the results, the main challenge is cleaning. After extracting just the body text of the articles, we saw that many of them did not mention Olympic legacy. Usually, Shine provides results where the words searched appear in peripheral locations of the webpage. Cleaning consists of removing all the information around the main text, such as images, adverts, menus and links. With the documents we needed in hand, we had to verify if their content is relevant or not to our analysis. Sometimes, the term Olympic legacy appears but it is not necessarily related to Rio and London Olympics or it is not the main topic of the article. The process of filtering demanded a huge effort of close reading to identify contexts. At the end, we have produced some charts to visualise word-trends and topics that pop up around legacy. Although the Shine search results are limited in terms of time - it searched up until 2013 - it has been very useful as an exploratory tool to conduct preliminary analysis in a small-scale, and to build web archive and web scraping methods before applying my methods to huge amounts of texts elsewhere. 

You can watch Caio de Castro Mello Santos & Daniela Cotta de Azevedo Major’s presentation on the EWA YouTube Channel.

*This project has received funding from the European Union’s Horizon 2020 research and innovation programme. For more information: cleopatra-project.eu.