MAKING CLIMATE CHANGE VISIBLE: TWITTER

Group members:

Sabine Niederer, Emile den Tex, Esther Hammelburg

Aim:

Describe how visual communication about climate change adopts a platform vernacular on Twitter.

Method

Question 1: For Twitter, what is the platform’s visual vernacular for the issue of climate change?

For the study of Twitter, we used the DMI-TCAT tool. The tool has collected climate change related tweets since 2012 (including tweets that mention climate, drought, flood, global warming, and globalwarming), adding up to a dataset of 134.402.719 tweets at the time of writing (30 June 2017 at noon). Due to the size of the data set, we created a separate dataset for a very narrow time frame, 24 May 2017 until 7 June 2017, a very eventful two weeks for the issue of climate change as it includes the day in which Donald Trump announced the USA would “exit the Paris Accord,” 1 June 2017. In addition, we narrowed the data to include only tweets mentioning “climate change” or climatechange (as text or hashtag).

• Query: “climate change” OR climatechange
• Date range: 24-05-2017 - 07-06-2017

The dataset then contains a total of 2.127.903 from 1.003.231 distinct users; 43.5% of the tweets contain links, 56.5% do not. The dataset is titled subset_climate_change and can be accessed through http://tcat.emiledentex.nl/analysis/ with the DMI-TCAT username and pw.

It is important to note that studying imagery through DMI-TCAT has limitations, as it is text-based research. If a tweet does not have any text mentioning the issue by name, it is not in the data set. See ‘Limitations’ for a further discussion hereof.

In DMI-TCAT, launch ‘Media frequency’, which produces a csv-file listing the media URLs and the number of times they have been tweeted and retweeted.

Note: this set contains only uploaded images, not the thumbnails that come with tweets that include links.

Question 2: Does recent imagery around the time of the US exit of the Paris Agreement (May-June 2017) differ from the imagery that was most engaged with on the platform one year prior (May-June 2016)?

Compare the 2017 data with data captured in the same two weeks, one year prior.

1. Query: “climate change” OR climatechange

2. Overall media frequency. Most shared media means content uploaded and shared by most users (retaining both original tweets and RTs).

3. Open image URLs and use Google Image Search (in Chrome) to retrieve the origins of the image

4. Save the top 10 images to a folder

5. Create a platform vernacular image

6. Close read the images.

Date range: 24-05-2016 - 07-06-2016

The dataset then contains a total of 418.111 tweets from 186.238 distinct users; 58.5% of the tweets contained links, 41.5% do not. The dataset is titled cc_nonfiltered_2016 and can be accessed through http://tcat7.digitalmethods.net/ with the DMI-TCAT username and pw.

Findings

The data set for Twitter consists of 20 images: 10 from May 24 - June 6 2017 and 10 from the same time period in 2016. Photographs, memes and infographics are represented in the data set, as well as stills from television or online video. About half of the images have text included. Regarding content in the photographs we see celebrities, politicians, and the polar bear. Landscapes and images of nature (aside from the polar bear) are noticeably absent. Images seem to follow current affairs pretty closely.

In the two eventful weeks in 2017, that included Trump visiting the Pope and the US pulling out of the Paris Climate agreement, we found that the top 10 images show different sides to the climate debate (program/anti-program discussion). The climate change images include a portrait of Barack Obama giving a climate change speech (the most-shared image), a UN climate action event, and polar bears (both in stylish b/w and starving on melting ice). The climate change skepticism imagery include Al Gore criticism related to his wealth and his large estate in California, and the Republican rep. Tim Walberg arguing that God will solve climate change.

In contrast, the top-shared images from 2016 do not include any skepticism, apart from a critical cartoon of Trump being skeptical (so skepticism-critique). The platform vernacular for climate change in 2016 mostly consisted of more archetypical climate campaigning imagery, such as floods, storms, polar bears. The only recurring image is that of a starving polar bear on floating ice.

Limitations

Limitations of the TCAT keyword search method are revealed by a comparison to ethnographic observations of climate change Twitter from the same time period (May 24-June 7, 2017). The US’s withdrawal from the Paris Agreement prompted a number of popular memes to emerge. However, some of these did not include “climate change” as text, so did not appear in the TCAT dataset. For example, French President Macron published a tweet containing no text, just a campaigning meme-like image which contained the slogan “Make Our Planet Great Again”. Clearly, this was a response to the US government’s decision, and a play on President Trump’s campaign slogan “Make America Great Again”. Interpreting the tweet as related to climate change is dependent not on a particular search term, but on knowledge of a number of contextual factors: the timing (around the Paris decision), the respective roles of Trump (climate change sceptic) and Macron (pro-climate action) and Trump’s 2016 campaign slogan. This is important as the Macron tweet was far more shared than any tweet within the TCAT database, with 240,323 Retweets and 396,172 Likes at the time of writing (30/6/17).

One potential triangulation method to address shortcoming of keyword searches is to collect additional tweets from key climate change Twitter users. A researcher with contextual knowledge of the field could set up a list of relevant Twitter users to monitor. In the case above, one might have anticipated that Macron would comment on any US action regarding the Paris Agreement, based on his own presidential campaign. However, another highly shared tweet came from a user one could not reasonably have anticipated. A tweet containing a humorous juxtaposition of memetic text (“Cracking open a cold one with the boys” and images (Trump with US officials and an iceberg) garnered 62,445 retweets at the time of writing. Again, this was far more highly shared than anything appearing in the TCAT ‘climate change’ keyword search, but was posted by a user with less than 1,000 followers and no history of tweeting about climate change. However, although this escaped detection using existing digital methods tools, it may be possible to devise methods in the future which could capture such tweets. As with the Macron tweet, there is no obviously relevant text within the tweet. However, the image content is more directly relevant and might be conducive to a method employing the Google Vision API, that could detect the image content as Trump and iceberg.

The success of these two tweets in comparison to those captured through the TCAT keyword search raises the possibility that tweets requiring more contextual information for interpretation are more liked than those that are more ‘literal’. In other words, tweet that have an ‘in-joke’ dimension may have higher potential for sharing.

-- NataliaSanchez - 17 Jul 2017