Visualising climate debate across the social web
Notes from the Making Climate Social Discovery Day
Social media has radically transformed the science communication landscape. It holds promise to democratise debate, and collapse ‘perceived’ boundaries between experts and non-experts. Despite growth in social media discussion about controversial topics like climate change, little is known about the nature of these interactions and how different social platforms play a part in shaping the debate.
Making Climate Social, funded through the ERSC Future Research Leaders program, is research led by Dr Warren Pearce (University of Sheffield) combining data mining and ethnographic methods to reveal new perspectives on climate change communication across social media.
I’m excited to be working with Warren to produce and publish a collection of visualisations that explore insightful questions asked of the project’s data. As a design producer and technologist, my job at the beginning of any project is to understand needs. This discovery process involves both understanding the project’s problem space, and understanding the needs of those who will use / engage with the project’s output.
Discovery Day
On 1 February 2017 — we brought together Making Climate Social’s advisory group for the first time. This ‘Discovery Day’ sought to gain diverse perspectives on state of the art for climate change and social media, and explore use cases for the project’s data analysis and visualisations.
Each member of the group was invited to respond to two questions, one broad and one more tailored to their domain expertise. This opening seminar helped us understand expert principles and insights in approaching projects of this nature. It also helped to map out possible future use cases for the project.
Consulting the Oracle
Mapping the questions that matter.
In any data driven project there is a tendency to treat the source data itself as the only starting point. By only focusing on what is immediately intuitable from known data, we risk not asking the questions that matter. I like to imagine data science as a process more akin to consulting an oracle than an almanac. In this way, the questions that can be framed are more expansive than statistical summaries and create opportunity for innovation. The afternoon kicked off with an exercise generating over 100 questions that were then clustered into emerging themes.
Below is a sample of themes, each with a question synthesised to represent an underlying set of questions:
- Images — What is the visual language of climate science / climate change on social media?
- The big picture — How does climate debate size up against the wider social web?
- Viral / Trending — Can we detect the triggers for climate conversation / debate?
- Automatic / Algorithmic — How much do algorithms and bots shape the climate debate?
- Attitudinal Change — Can we know whether discussing climate science online changes attitudes, minds and policy?
- Platform Specific Issues — How do platforms shape what can be said?
- Platform Movement — As private messenger based social networks grow, how might we understand their role in the climate communication landscape?
- Demographics — Who are the people joining the conversation?
Hypothesis Driven Development — Sort Of
When working with software teams, I draw on agile software development principles to handle the inherent uncertainty of building something new. Agile software is built in short iterative cycles, rather sequentially managed stages bookended with sign offs. The currency of agile software development is the user story. This is a structured statement with the form:
<A descriptive title>
As a <type of user> I want <feature/functionality> so that I can <goal>.
User stories, usually come from user and market research. They are also completed with acceptance criteria — tests and metrics to validate expected outcomes. This approach helps align conversations, about the system, from the view point of the user. An extension of agile implements stories and their criteria as automated tests written in code, an approach called Test Driven Development (TDD).
However, how how do you imagine these stories or write tests for features not yet defined? As Making Climate Social sits within an academic context, I thought that a cousin of TDD, Hypothesis Driven Development, might be a useful framework to discuss possible use cases and features. Hypothesis Driven Development recasts user stories as experiments. Reflecting on ideas, questions, and themes raised earlier in the day, we invited delegates to produce a short proposal for an ‘experiment’ that would be of value to them.
Turning needs into meaningful features
Outlined below are three core features for Making Climate Social’s data visualisations that emerged from the Discovery Day. There was a clear need for analysis across different social platforms. Many projects researching social media and science communication focus predominantly on Twitter. Consider: Is Reddit used more as platform for breaking scientific discoveries than Twitter? For example: NASA’s recent use of Reddit’s Ask Me Anything (AMA) to discuss Exoplanets. Could performing analysis across platforms offer better insights and support the project’s aim to understand climate communication in context?
Content
What are the key topics being discussed? How are topics connected to each other? How are topics shaped over time? Who contributes to certain topics? What is the visual language of the climate change debate?
Pathways
How do people, topics, and media move across the social web over time? What is the life span of a shared links / media? Where are scientific articles incorporated into the conversation?
Triggers
What are the significant statements / events / content that trigger discussion and debate? How might we model the ripple effects of significant statements across social platforms?
Over the next 6 months we will be bringing together a creative software team to analyse data, and prototype ways visualise the things we find. We will be publishing regular notes, thoughts and work in progress here. We would really welcome your feedback.
What questions about climate change on social media would you want to ask?
