‘Algorithmic play’ artefact – ‘Algorithmic systems: entanglements of human-machinic relations’

‘These algorithmic systems are not standalone little boxes, but massive, networked ones with hundreds of hands reaching into them, tweaking and tuning, swapping out parts and experimenting with new arrangements…we need to examine the logic that guides the hands…’ (Seaver 2013: 10)

Early on this block, I played with a range of different algorithms, looking at the recommendations served up by my existing activity. I have created a “recommendations feed” for my lifestream, and also made general notes and screenshots in number of posts.

Later on, and in thinking specifically about digital education, and tying the activity with our MOOC exploration, I first looked at my FutureLearn recommendations, and finally, focused on Coursera – where I was able to provide false data for my “profile” and “learning plan” and see the alterations in recommended courses.

Some initial conclusions

Many of the “recommendation engines” I played with, such as SoundCloud, Spotify and Twitter, led me into what I perceived to be a “you loop“, “filter bubble” or “echo chamber”. Google’s autocomplete showed some signs of reproducing existing biases, perhaps amplifying dominant views held by other users. Google may also be making assumptions based on data they hold about me, which may have skewed YouTube search results, although it would be interesting to compare results with other users or in different locations, an approach Kitchin (2017: 21) discusses. I have written a little about ethical concerns and my Google data also.

Moving my focus to Coursera, in its recommendations they appeared to privilege courses with a information technology/computer science focus, although the range of available courses is clearly a factor too. In any case, the founders’ background in computer science, and education at a Western university, shone through despite attempts to tweak my “profile” and “learning plan” (I have written a little about the “profile” ethical issues also). This appears to be a common theme in “apps” or websites” developed by Western companies, and the design processes utilised (whereby products are developed first for the “profile” of the founder(s), and secondarily for others) arguably creates exclusions for some (and a strong sense of “this wasn’t designed for me”) and inclusions for others (notably switching my profile to “software engineer” produced a wealth of “relevant“ results which I could tailor to my choosing).

My artefact: Algorithmic systems: entanglements of human-machinic relations

You can see the Coursera ‘algorithmic play’ posts in a lifestream feed, and also through this CodePen (which brings in the same feed but presents, ranks and orders it slightly differently). How might consuming the feeds through different ranking systems affect your perception, and how might the pre-amble surrounding it have changed your conclusions and what you click on?

View the CodePen

‘Algorithmic systems: entanglements of human-machinic relations’
‘Algorithmic systems: entanglements of human-machinic relations’

You can also access the “code behind” the CodePen. However, even though access is granted to this code, it will not necessarily be easy to make sense of it (Kitchin 2017: 20; Seaver 2013). While presented as ‘open’ source, is it really ‘open’ and who for? What exclusions might this create?

The CodePen (while slightly misleadingly called an “Educational Algorithm”) is a very basic attempt to present the feed of Coursera recommendations in a different fashion for comparison, while provoking some thought around algorithmic systems in general. There is no complex processing, just a little reordering. It also does not store any data (information entered is used within the browser to rank articles, but not stored in an external database), and the text descriptions are fictional – it is just for visual/demonstration purposes.

Reflections on the CodePen

Inspired by Kitchin (2017: 23), I have written a few words reflecting on my experiences, while acknowledging the limitations of the subjectiveness of this approach.

The CodePen was adapted (“forked“) from an existing CodePen, which provided the majority of the base upon which (with my very limited skills!) I could tweak very slightly to add a basic feed of the Coursera recommendations by pasting in another bit of code (which turned out to be old and broken) and then ultimately another bit of example code (this is from a Google-“led” project). It is very much a case of different bits of code “hacked” together, full of little “bugs” and “glitches” and not particularly well designed or written! I was keen to strictly limit the time spent on it, although I know much time could be spent tidying and refining it.

Presumably, similar time constraints, albeit with more resources/testing, affect development of, say, Facebook/Google etc. algorithms and lead to mistakes. After all, Facebook’s internal motto used to be ‘move fast and break things’, although this race to create a “minimum viable product” and “disrupt” (regardless of the consequences) is increasingly criticised.

In any case, this CodePen is like a snapshot of an experimental “work in progress” (which others are welcome to “fork”, use and adapt) and brings to mind the messy, ontogenetic, emergent, and non-static nature of algorithmic systems (Kitchin 2017: 20-21).

Ultimately, my aim is to raise some questions about some details of this process and, since most of the code was “stuck together” from bits of free and ‘open’ source code, how the culture of the individuals/teams is significant too. As Seaver (2013: 10) puts it…

‘…when our object of interest is the algorithmic system, “cultural” details are technical details — the tendencies of an engineering team are as significant as the tendencies of a sorting algorithm…’

…and, given that a large portion of the code came from a Google-led project (and Twitter too), how might the tendencies of those teams have created inclusions/exclusions? Furthermore, as the focus of my artefact is Coursera, whose founders both have experience working at Stanford University and Google, how might the tendencies there have guided Coursera and, subsequently, my CodePen?

Finally, given that Coursera presents itself as a universal educational space, where its vision is ‘a world where anyone, anywhere can transform their life by accessing the world’s best learning experience’, what are the implications of this for digital education? In my initial explorations, my perceptions are that computer science and information technology disciplines from Western universities are prioritised by Coursera, in general and through their algorithmic systems. However, further research is needed to, as Kitchin (2017: 25) puts it, ‘unpack the full socio-technical assemblage’.

View references

Data collected through Coursera profile – what are the ethical issues at stake?

The data you can enter through the Coursera “profile”, while it is not compulsory, promises potential career opportunities, new connections and course recommendations. This may be of less interest to those who have an existing level of education, a job and existing connections; however, if you are visiting Coursera with the hope of increased career opportunities, it may be significant (although it is difficult to tell without further analysis or Coursera data).

A quick check of the Coursera privacy policy reveals that ‘general course data’ and site ‘activity’ may be shared with ‘Content Providers and other business partners’, including personally identifiable information, and ‘Content Providers and other business partners may share information about their products and services that may be of interest to you where they are legally entitled to do so’.

In addition to the site activity data (presumably course searches, enrolments and so on) collected by Coursera, the additional information you can provide to personalise your ‘learning experience’ and recommendations is fairly extensive, including work experience, education, career goals, location, age and gender:

Coursera profile Coursera profile

While it is possible through this page to limit the information to ‘only me’, ‘the Coursera community’ or ‘everyone on the web’, presumably the privacy policy will still allow Coursera staff, and associated ‘content providers’ or ‘business partners’ to access and analyse this data. The options presented (for which I selected ‘only me’ in all cases) give a slightly false and misleading sense of privacy, since the privacy policy outlines that it should say ‘only me, plus Coursera staff, Content Providers and business partners’ – that is, assuming I have understood the policy correctly, although presumably many will never read it at all.

There do not seem to be any options (easily visible on this page, at least) for hiding all your data from everyone else (including Coursera staff, ‘Content Providers’ and ‘business partners’), nor do there appear to be any options for customising who can view your site activity. For an educational site – where some may be following a promise of improved career opportunities, and where everyone is not beginning at the same ‘starting point’ – it, in my opinion, seems appropriate to be able to hide all data from everyone.

Viewing different Coursera courses to influence my recommendations

The tweaks to my Coursera profile and learning plan have had a fairly limited effect so far on my Coursera recommendations.

I notice my recently viewed courses have an impact, so will look to alter this:

Recently viewed courses in Coursera
Recently viewed courses in Coursera

First, I am switching my profile and learning plan to a nurse in the healthcare industry:

Coursera profile
Coursera profile
Coursera learning plan
Coursera learning plan

…courses that others identifying themselves as nurses now appear…

Coursera - 'People who are Nurses took these courses'
Coursera – ‘People who are Nurses took these courses’

Notably, the first is “nursing informatics” – could this be another example of information technology dominating results?

I view some courses related to ‘Everyday Parenting’, ‘Mindfulness’, ‘Well-Being’ and ‘Buddhism and Modern Psychology’ and ‘Social Psychology’.

Below some more computer science/information technology degree recommendations…

Coursera 'Earn Your Degree'
Coursera ‘Earn Your Degree’

There are some courses displayed on ‘Personal Development’. Many are not particularly related to the areas I specified, however it is a rare opportunity to see recommended courses that are not computer science or information technology.

Coursera Personal Development
Coursera Personal Development

My explorations seem to again show a privilege towards computer science subjects – again, not surprising given the background of the founders.

However, this limited focus does seem slightly at odds with Coursera’s own slogan:

‘We envision a world where anyone, anywhere can transform their life by accessing the world’s best learning experience.’
(About Coursera)

As previously discussed, the approach for those from a Western university-educated computer science background to build something for themselves, raise funds through investment but then market it as a “universal” solution that is “best” for all appears quite common.

Tweaking my “profile” in Coursera to “software engineer”

Further to setting my initial (false) “profile” and playing with my Coursera “learning plan”, I have now tweaked my profile to indicate I am a software engineer at “Executive Level” at Facebook, with a masters :

Tweaking my Coursera profile
Tweaking my Coursera profile

I have also set my learning plan so that I am a “software engineer” in the “technology” industry:

Coursera learning plan
Coursera learning plan

The key difference here is the recommended course list, which now suggests courses that other software engineers have taken:

Coursera recommendations
Coursera recommendations

There seem to be a wealth of courses in the area, which is perhaps unsurprising given my other experiences of the site so far.

SoundCloud and Spotify recommendations – a “you loop”?

Here is the Spotify playlist which has been connected to my lifestream…

EDC Spotify playlist
EDC Spotify playlist

…and today’s recommended songs…

Spotify recommended songs
Spotify recommended songs

…which, with a few exceptions, are largely very “similar” or songs from the same albums.

My SoundCloud recommendations appear to be partly influenced by listening to a podcast from Meet The Education Researcher

SoundCloud 'Artists You Should Know'
SoundCloud ‘Artists You Should Know’
SoundCloud 'Artists You Should Know'
SoundCloud ‘Artists You Should Know’


Are these examples of the algorithm pushing “similar” content and perhaps also changing my perception of what I should listen to? Have I been in a “you loop“? Have my recommendations been influenced by others listening to them?

How is a YouTube search for ‘algorithms in education’ altered when I am signed in?

Below are the results of a YouTube search for ‘algorithms in education’ – on the left I am signed in, on the right I am not.

YouTube search results - comparing the difference when signed in
YouTube search results for ‘algorithms in education’ – comparing the difference when signed in

The results are subtly different – when I am signed in, slightly more ‘advanced’ videos about algorithms (one from Harvard University) are displayed. Perhaps this is due to information Google holds on my age and education, or due to the fact I have watched and liked a number of longer university lectures and interviews on this course.

This speaks to the way algorithms are ‘ontogenetic, performative and contingent’ (Kitchin 2017: 21) – they are not static nor fixed, vary from user to user, from location to location and can often involve randomness.

Changing my “learning plan” in Coursera

I changed my Coursera “learning plan” to indicate that I am a Teacher/Tutor/Instructor in the Education industry, to compare the results with my previous exploration of Coursera.

The results are more varied (and not exclusively focused on software development or the “tech” industry), however there are still various programming, data/computer science and business options presented (despite expressing no preference for this kind of industry):

Coursera recommendations
Coursera recommendations after altering my “learning plan”

Coursera recommendations (based on false data) – what inclusions and exclusions are apparent?

I am experimenting with inputting false information about myself in Coursera, in order to see the difference in algorithmic recommendations. Here is how I described myself…

False data provided to Coursera
False data provided to Coursera

… and here are some recommendations provided after entering the above data…

Recommendations provided by Coursera
Recommendations provided by Coursera

The top listed courses are exclusively technology-based and “offered by” Google, and appear to have no direct connection to my listed industry “Health and Medicine”…

While my explorations were very limited here, in some ways this seems fairly consistent with my experiences of using certain (but not all) MOOC or educational course/video sites (and even more general “apps”). As soon as you step outside of the area of computer science, the range of courses is more limited, despite the sites themselves being presented as general educational sites. In looking to change my “learning plan” options (which change your profile and recommendations) revealed the “default” or “suggested” text, presented before you enter your own profile options:

Setting your Coursera "learning plan"
Setting your Coursera “learning plan”

You can see the results of my profile/”learning plan” alterations here. However, at this stage of deciding my profile options, the “software engineer” who works in “tech” seems to be the “default” starting point here. This is all perhaps no surprise given that Coursera was set up by Stanford computer scientists; as often seems the way, the developers build something for themselves (ensuring a seamless user experience for their own circumstances) and then only later branch out.

One example outside of education here is the online bank Monzo, whose early customer base was ‘95% male, 100% iPhone-owning, and highly concentrated in London’ (Guardian 2019). This description mirrors the co-founder Tom Blomfield, as he himself admits:

‘Our early customer was male, they lived in London, they were 31 years old, they had an iPhone and worked in technology. They were me. I’ve just described myself. Which has huge advantages, right? It’s very easy to know what I want.’ (The Finanser 2017)

While Monzo does claim to have a focus on social inclusion (This is Money 2019), why is this always seemingly secondary to building the app, gaining users (similar to themselves) and getting investors on board? Should social inclusion, whereby apps are designed for all users in a democratic fashion where everyone has a say, not be inherent in the very beginning planning, design and development processes? There may be a place here for considering platform cooperativism, inclusive codesign and participatory design approaches here (see Beck 2002; Scholz and Schneider 2016; West-Puckett et al. 2018).

Coming back to education, if Coursera have taken a similar approach as Monzo to designing their platform and building up their catalogue of courses, it is perhaps concerning that who do not mirror the designers and developers may be left excluded and on the margins.

Conversely, an inclusive codesign approach may have produced different results. As Trebor Scholz (P2P Foundation 2017) explains:

‘The importance of inclusive codesign has been one of the central insights for us. Codesign is the opposite of masculine Silicon Valley “waterfall model of software design,” which means that you build a platform and then reach out to potential users. We follow a more feminine approach to building platforms where the people who are meant to populate the platform are part of building it from the very first day. We also design for outliers: disabled people and other people on the margins who don’t fit into the cookie-cutter notions of software design of Silicon Valley.’
Trebor Scholz (P2P Foundation 2017)