Community formation that transcends politics in the age of filter bubbles – an example from Facebook

By | Scholarly blog post | No Comments

The recent fires in the Christchurch Port Hills caused enormous destruction and disturbance to those living in the area. Tragically a life was lost fighting the blaze, and there was major disruption when about 450 houses were evacuated, 11 homes destroyed, and over 1800 hectares scorched in the region between the suburbs of Cashmere, Westmorland, Kennedy’s Bush and Governor’s Bay.

Ross Younger Creative Commons Attribution-ShareAlike 2.0 Generic

Concerned citizens during the crisis created a Facebook group named “Evacuation housing” where people could share information, and offer food or shelter to those who had been evacuated. I wanted to understand who participates in such social media responses to a crisis, so decided to study the political affiliations of the people who joined this group. I did this by examining publicly available information about each member’s past behaviour on Facebook to determine their political orientation, and then plotted this data on several graphs to help visualise the findings.


Fig 1.0. This graph shows the aggregated political orientation of everyone in the group. Political orientation is determined by the individual users like-history.


Fig 1.1. This graph shows the aggregated political orientation of everyone in the group relative to the country average on Facebook. It shows political groups that are either over- or under-represented in the group.

Fig. 1.2. This network diagram shows all the interactions between individual users who are part of the Facebook group “Evacuation housing”. The color of the nodes correspond to the same political affiliation as in fig. 1.0 and fig 1.1 Names are omitted from the nodes in the network in order to preserve privacy.


A quick look at the method

The method behind the study is simple, though not easily implementable. First, data was collected using the public Facebook Graph API 2.7 I collected data on all individual users who had been active on any of the public Facebook pages represented by a politician currently in parliament. All users who had not liked at least 7 posts from any of the pages were filtered out, which produces a little less than 200,000 unique users in New Zealand. Each user’s political affiliation is based on the political party from which the users has liked to most post on a percentage basis. Almost 1100 users could be identified with a probable political affiliation out of the 3500 members of the public group “Evacuation Housing”


At the outset Fig 1.0 shows a fairly balanced distribution of affiliation with the political parties. On Fig 1.1 we see that Green and Labour are somewhat over-represented. Fig 1.2 shows that interactions within the group do not form clusters of users who are only affiliated with the same party.

In the age of filter bubbles, echo-chambers and behaviour-altering algorithms we see that online spaces still allow different people to come together and help each other in a time of crisis. This very small, very limited data-driven look into a spontaneously created Facebook group shows how political orientation is not overly skewed toward one particular group when it comes to issue based community formation. Even though some political groups are overrepresented it is not enough to constitute a so-called echo chamber, a place where only likeminded people interact. The actual significance of over- and underrepresentation by political groups in this case will require deeper analysis.
This study is also a peek into how the public behaviour of individual users can be applied to create segments and find out how these segments are mobilized for a certain cause. Future studies could involve the same political segmentation to be used to in other contexts or other segments could be created using the same basic method.

Santa-fy Yourself

By | Uncategorized | No Comments

For the last three years, the UC Arts Digital Lab has been sharing a floor with the College of Arts Office staff and has been invited to enter their Christmas-door decorating competition. This is an invitation that we do not take lightly, mostly due to an over-zealous thirst for competition, but also because we love decorating things, and (of course) we love Christmas.

This year, we decided that as a Digital Lab our door should reflect the digital skills that we have on offer. So armed with Python and a webcam, we began the process of creating a script that would turn people walking down the hallway into Santa! The idea was to detect people’s faces using facial recognition software and to then to superimpose a beard and a hat onto their face and project it onto a screen.

Read More

Eric Meyer visit

By | Events, Seminars | No Comments

We were delighted to welcome Professor Eric Meyer back to the University of Canterbury for a brief visit over the past two days. Eric previously visited the UC Digital Humanities programme as a Visiting Canterbury Fellow in 2014, and was back in New Zealand to speak at the National Digital Forum in Wellington earlier this week. He very kindly added a Christchurch leg to his trip, so we were able to hear him speak yesterday about work arising from his book Knowledge Machines (with Ralph Schroeder, published in 2015 by The MIT Press), then this morning we attended his workshop entitled “Metrics and Measurement: The Impacts of Digital Resources and Collections”.

The workshop stimulated some broad ranging discussion that moved from ways of measuring impact into exploration of data collection tools and methods of data analysis in various fields of research. A number of our Media & Communications postgrads took the opportunity to ask Eric about possibilities and limitations of using social media data, which led to some interesting debate.

Thank you, Eric & Michelle, for making the trip down to see us in Christchurch – we wish you safe travels for the journey home. And thanks National Digital Forum for inviting Eric to speak and getting him over to our part of the world!


Eric Meyer speaking at the University of Canterbury on 24 November 2016.

Professor Eric Meyer of the Oxford Internet Institute addresses UC staff and postgrads

Metrics and Measurement: The Impacts of Digital Resources and Collections

By | Events, Visitors | No Comments

Following his talk on Thursday 24 November, Professor Eric Meyer will also lead a workshop while he is at UC:

Metrics and Measurement: The Impacts of Digital Resources and Collections
Friday 25 November, 9am-12pm, Macmillan Brown PS 208

This workshop will present a framework and best practices for measuring usage and impact of digitised scholarly resources. The workshop will cover quantitative and qualitative methods outlined in the Toolkit for the Impact of Digitised Scholarly Resources and how organisations can apply these to their own collections and projects. Participants will learn from case studies and work through participant-provided examples to gain a better understanding of:

  • Web presence, what it means, and how it can be measured with analytics;
  • Social media data, and how one can get it and use it for understanding impacts both quantitatively and qualitatively;
  • Scientometric data, and how one can interpret it;
  • Interviewing and surveying users.

Morning tea will be provided. Please RSVP by Monday 21 October to for catering purposes

To get the most out of the workshop, participants will need to bring a laptop that can connect to the Internet.

Eric Meyer is Professor of Social Informatics and Director of Graduate Studies at the Oxford Internet Institute, a multidisciplinary department at the University of Oxford which undertakes teaching and research focused on understanding life online. Eric’s work focuses on shifts in work, knowledge creation, and human interactions when digital technologies replace previously non-digital counterparts. His research in this area has included studies of the impacts of digital collections in libraries and museums, digital practices in the arts, the use of digital images in biology, and digital information practices in the sciences and humanities.