Much of the early stages of the blog, and continuing (though less so) throughout the course, I found were spend engaging with the limits of the technologies I wanted to use. Spam, IFTTT, issues with tags and categories, the design of the blog, the issues of trying to embed Storify, were all challenges that were publicly wrestled with in the lifestream. I also sometimes had issues with synching and cloud storage–I blogged from my laptop at home, but often took notes or photographs on my phone, tablet or other computer, and glitches in synching could mean that trying to upload the images or notes to the blog was delayed.

This wrestling surprised me, because I have been using most of these technologies in my leisure time, or for work, for some time. I think the pressure to blog nearly every day while only having a day or two to focus on my studies meant I was always in a rush, so any lag or glitch could throw the whole thing off.  After semester started in my day job, too, I found it difficult to spend evenings and weekends on the course, since I often had to be at work. I had expected to be able to vary my working days to counteract that, but that was not possible for structural reasons, and that was a challenge.

In the first few weeks, there was too much Twitter and not enough of anything else. I often discuss learning, teaching and technology on that ‘academic Twitter’, so the #mscedc hashtag was a natural extension of those conversations, and fitted comfortably with the people I already follow and the people who already follow me. During the course, I gained nearly 100 followers overall, suggesting that I didn’t alienate most of my audience, and that they appreciated the contributions I was making to their streams.

On the other hand, my Instagram community is small, it’s made up of people I personally know, and they share pictures of their lunch, of their holidays, of their homes. It’s a leisured and personal space, so sharing technology news or reflections on the digital would be less appropriate. I eventually worked out that I could use it to document my own embodied working practices. I could photograph my experiences blogging with a bowl of pie, or reading in a café. These would contribute to my community but also to the lifestream.

I was often unsure about the value of the lifestream as a public blog. The learning journal blog focuses on a very small audience, primarily myself, my tutor, and sometimes the other members of the course. The format of the lifestream is difficult for an outsider to navigate or read.

Instead, I longed to craft things from what we’ve learned in the course to share with the world. I enjoyed live-blogging the readings, and continue to find that one of the most useful ways to engage with the course, the medium and the content. I wonder if this might be a valuable way of thinking about learning in public in the future.

It’s autumn here. Warm during the day, cooler in the evenings. We are just beginning our academic year. We have not yet ‘fallen back’ out of daylight savings time. Scotland is a long way away.

To be ‘with’ the class synchronously, I need to get up very early, or stay up late. But my job requires me to get up early, and stay up late at this time of year.

On Thursday, instead of being at the Hangout, I was trying to complete the tutorial timetable, interview and place new tutors, and do spot checks to check all the classes were running smoothly.

On Friday, instead of being at the Hangout, I was teaching at ArtsWrite, an intensive doctoral writing weekend at the University of Melbourne. (This is a picture of the stamps to reward attendees for meeting their goals).

I wonder if my embodied reality here on the other side of the world is too different, too far away, too out of synch with Edinburgh. I wonder if digital learning can overcome those distances and disjuncts. I wonder if it should.

This year, my job has little to do with online learning. This year, the questions I find raised by the readings make me more critical of the university, rather than more hopeful. This year, the sun is shining, and I am inside on my computer. I wonder if I’ve got that wrong.

Since CAPTCHA was put on the blog, I only had 19 spam comments.

My Storify of tweets in week 10:

https://storify.com/katrinafee/mscedc-week-10

The view that data and analytics offer a new mode of thinking and a new model of discovery is at least partially rooted in the artificial intelligence and machine learning fields. Halevy, Norvig, and Pereira (2009) argue for the “unreasonable effectiveness of data” (p. 8), stating that machine learning and analytics can help computers to tackle even the most challenging knowledge tasks, such as understanding human language. Hey, Tansley, and Tolle (2009) are more bold in their assertions, arguing that data analytics represent the emergence of a new approach to science.(p. 1381-2).

Here we have a connection between the android/cyborg of section one, the post-human ethnography of section 2, and the algorithms of section 3.

There are many fields that Siemens covers, but the developments described in E-learning are less speculative, they are happening now in Universities.

E-learning: The growth of online learning, particularly in higher education (T. Anderson, 2008; Andrews & Haythornthwaite, 2007; Haythornthwaite & Andrews, 2011), has contributed to the advancement of LA as student data can be captured and made available for analysis. When learners use an LMS, social media, or similar online tools, their clicks, navigation patterns, time on task, social networks, information flow, and concept development through discus- sions can be tracked. The rapid development of massive open online courses offers additional data for researchers to evaluate teaching and learning in online environments (Chronicle of Higher Education, 2012). (p.1384).

I personally have an issue with using learning analytics to tell us anything except about the interface with machine learning. We can look at our data, considering bounce rates, load times, click throughs etc, and we can decide that students are using the pages as intended… or not. But we really can’t know anything about how much students learn or understand. Even if we test students, we can only know how students respond to the tests—from my own experience, this is often more useful to demonstrate that I designed the test badly, or phrased the questions ambiguously, than to demonstrate student’s understanding or recall. (Frequent small stakes questions are very useful to prompt students to recall information and reproduce it, helping them to reflect and aiding memory. But it’s the actions and reactions prompted by the learning interface but taking place outside of it, that really matters.)

Siemens also notes that this kind of data is “academic analytics” not “research challenges in learning”, that is, it “involved the adoption of business intelligence (BI) to the academic sector (Goldstein, 2005).” For this reason “Commercial tools are the most developed… Research and open analytics tools are not as developed” (p. 1836).

Siemens (2013) Fig 1, p. 1837.

Siemens reminds us that algorithmic or machine data is more useful when it is supporting “human effort”:

To be effective, holistic, and transferable, future analytics projects must afford the capacity to include additional data through observation and human manipulation of the existing data sets. (p. 3987)

I was horrified to read:

Curriculum in schools and higher education is generally preplanned. Designers create course content, interaction, and support resources well before any learner arrives in a course (online or on campus). Through the use of analytics, educational institutions can restructure learning design processes. When learning designers have access to information about learner success following a tutorial or the impact of explanatory text on student performance during assessment, they can incorporate that feedback into future design of learning content.

*begin rant*

Online teaching and blended learning requires hours of preparation that makes it hard to change a class mid-lesson. I now lecture without powerpoint, or with only images if possible, as then I can read the room, take questions, realise the lecture is not meeting the student’s needs, and draw on my knowledge to re-form the teaching content. In small group teaching, I do the same. One semester I taught four tutorials in a row: I would plan a single lesson, but each time I delivered it, I would amend, react, expand or contract parts of the class. I sometimes gave four radically different classes (because I wasn’t delivering pre-planned content, but the group and I were exploring the content together).

For this reason, the ‘personalisation’ of student learning is also problematic. Human beings are social animals, and we learn as a social interaction. This means that personalisation alone is insufficient for effective learning. We need to be able to alter teaching to support groups, and realise that students will behave differently in different cohorts. The personality of the teacher has been demonstrated to have an enormous influence on student engagement (Shelvin et al 2000, Williams and Ceci 1997, Murray, Rushton, & Paunonen, 1990), and the personality of other students in a class is also significant. While there are ‘lone wolf’ students, this does not invalidate the need to explore group dynamics, but rather demonstrated that there are multiple factors that should be explored in learning analytics for them to be useful for learners.

An empathetic human teacher doesn’t need to wait till after a tutorial to incorporate feedback into the design of learning content, but can (and should) be redesigning the learning content in the classroom. Every human teacher should be looking at “the impact of explanatory text on student performance during assessment” as they mark, and this should be, as I understand teaching (and have been doing for the last decade), incorporated into the next iteration. The question that 30% of students clearly misunderstood? The LMS quiz that was aced by native English speakers and tanked by international students? These things we are already quantifying without big data and without E-learning. In fact, we’re waiting for E-learning to catch up with the speed that an experienced teacher can manage with a whiteboard and marker, or document camera (another modern technology).

*rant the second*

More importantly,

Concerns about data quality, sufficient scope of the data captured to reflect accurately the learning experience, privacy, and ethics of analytics are among the most significant concerns (see Slade & Prinsloo, 2013). (p. 1392)

Privacy is, I think, one of the most significant issues with modern learning analytics. I used to teach a class that required the production of a weekly ‘learning journal’. However, the journal was only viewed twice in the semester. If a student decided to write in larger chunks, or individually, if they journalled before or after class, if they did it all in a big go just before handing it in… I didn’t know. I could tell the students who did it badly—and it was designed to help students reflect and engage as they went. But students had privacy to produce assessment at their own pace.

This meant that students didn’t have to tell me they were sick, they were busy at work, that their children were home from school with chicken pox. They did have to attend my tutorials, but that was 1 hour a week. Other than that, they could learn privately.

I work in blocks and then distribute my learning across the week though scheduling and back dating. I found IFTTT meant that learning intruded too far into my private space. I had to act to suit the assessment and the algorithm, rather than learn. I chose to learn.

A public blog means that I have to be public about my health, my busyness, my business. Or at least the markers of those interruptions are publicly available.

Siemens quotes two scholars writing in 1964, 50 years later, their concerns remain pertinent:

Ellul (1964) stated that technique and technical processes strive for the “mechanization of everything it encounters” (p. 12).
“Most of our institutions of higher learning are as thoroughly automated as a modern steel plant” (Mumford, 1964, p. 274). (p. 1395)

Seimens suggests:

The learning process is creative, requiring the generation of new ideas, approaches, and concepts. Analytics, in contrast, is about identifying and revealing what already exists. (p. 1395)

Ellul and Mumford might also remind us that the purpose of the mechanised plant is to repeat the same process over and over again, as quickly and consistently as possible. Humans, as I have argued above, are not very good at being consistent, and as Adorno and Foucault might argue, any attempt to make students into machine-goods is an act of oppression. What’s more, Foucault would suggest that the assessing gaze of the teacher is another form of coercion.

Unlike Facebook or Amazon, who track our clicks, absences and contributions to give us more of what we want, teachers judge our clicks, absenses and contributions against a regime and rubric of value judgments.

*rant trails off, need to go make tea*

In my reflection on week 9, I wrote:

I managed a couple of tweets in the tweetorial (and neither of them had space for the hashtag, #fail).

The tweet archive, however, suggested that across that week, had contributed 15 tweets, and was the 8th most active user.

Active users 3 March-15 March 2015

Now, the top engagers tweeted four times as much as me, and my tweets would not have contributed to the tweet spike around the tweetorial…

Tweets over time, showing a spike for the tweetorial

I was also fascinated to see that most people were using the Twitter desktop client, and then TweetDeck. As I’ve been exploring recently, almost all my Twitter use is via the mobile app. It is possible that almost half of all iPhone tweets, therefore, were mine.

What these visualisations do not enable us to do, though, is judge value, only quantity. Moreover, there is no differentiation between the joking conversations about the spam (deer antlers, Michael Kors handbags) [sociality value], the sharing of resources [research value], the discussion around content [analysis value].

It’s a pretty blunt instrument. I loved Claire’s reponse:

This is what it feels like after looking at Tweet Archivist and Keyhole. #mscedc pic.twitter.com/vVzwgwU4Oq

— Clare Hampton (@clarehampton) March 19, 2015

***

Updated 20 March:

The Keyhole analytics are worse. I don’t have geolocation for Twitter enabled, so I have literally been wiped off the map (zero engagement from Australia!).

The gender engagement seemed off too, so I went and did some analysis of my own. My analysis is that 43% of tweets were from female participants, and that 36% of the participants are female. I’m not sure how Keyhole identifies gender (I tried to search for how Keyhole counts gender, but I couldn’t work it out. Twitter does not require gender for registration). But the algorithm is clearly miscounting, and therefore dangerously misrepresents our engagement. To go from over 1/3 to 1/10? That’s ridiculous.

Algorithmic counting can be reductive, but it is clearly also erasing certain identites, without being explicit about how it makes it’s judgements. Even assuming that algorithms are rough approximations can be extraordinarily problematic, therefore.

"Did QuillConnect tell you anything you didn't know?" @katrinafee I felt it tried to algorithmically 'normalise' my social media use #mscedc

— Sian Bayne (@sbayne) March 6, 2015

Reading PJ’s blog, however, inspired me to go back and try again.

***

To start, some numbers:

You have been a Twitter user for four years and you tweet more than most of your followers. You post 66 tweets a week while your followers average 7 per week.  Further, you have 916 followers listening to you … You are in the 96th percentile of Twitter users measured by followers.

During MScEDC I’ve been tweeting more than normal–this week I tweeted 86 times, last week I tweeted over 150 times. The numbers are correct, as far as I can measure them. I was surprised that someone with less than 1000 followers was so influential, but it would explain why some accounts with just over 1000 followers were rated influential enough to appear alongside major platforms like The New York Times or The Economist on my Discovery timeline.

I have manually created my own ‘youloop’–most of the people I follow and those who follow me are interested in the same topics as me: “Your important topics match those most tweeted about by followers who are similar to you.” “The hashtags most often used by your followers similar to you have been #mscedc, #phdchat, and #talkhe.” This is unsurprising!

Second, issues of catagorisation.  Some of my catagories are much more specific than QuillConnect’s (History, Writing, Academia probably all count as ‘Education’). I made no distinction between sources of interest, where QuillConnect lists ‘Entertainment, Arts, Music, Television, Celebrity’. On the other hand, I make a distinction between ‘Geek culture’ and ‘Popular culture’ and ‘Writing’. QuillConnect does not list ‘Food’ as a catagory, which is surprising (though perhaps less so on Twitter than Instagram or Facebook).

Third, issues of sentiment.  QuillConnect is interested in ‘positive’, ‘negative’ or neutral’ language. I listed content analysis of genre. However, it would not be hard to suggest that Diversity, Hard News, Online Security are likely to skew negative, where as Humour, Inspiration, Visually attractive are likely to skew positive.

QuillConnect analysis of my tweets

I therefore recoded my tweets to try to ‘reverse engineer’ my Discovery timeline analysis. I then compared them to the QuillAnalysis report.

I mostly tweet about politics (8%) according to QuillConnect. From my own content analysis, Twitter reads this as more likely to mean ‘diversity politics’ than ‘hard news politics’.

I regularly tweet about Education, and education tweets are most likely to be served to me in my Discovery timeline (both mobile and desktop).

Twitter thinks I’m more interested in Technology than Science, while QuillConnect thinks it’s the other way around. (I tend to agree with Twitter).

I tweet less than most people about Entertainment, Television, Music and Celebrity. As I don’t own a television and films give me migraines, this is not surprising. Discovery tends to agree–my tweets that did reference television tended to do so obliquely. A gif referencing Star Trek was really about technology, a tweet mentioning television marathons was really a joke about tenure (education), a tweet about HBO was more about structures of technology companies than the content of any shows.

I tweet about ‘Arts’ according to QuillConnect.  Discovery suggests that this means, sculpture, historical artifacts, photography, literature.

My Twitter is regarded as ‘neutral’ by QuillConnect, and “for #mscedc, tweets containing the hashtag are predominantly neutral in tone.” When I re-coded my Discovery timeline, I counted 6 positive and 7 negative tweets on mobile (some of the tweets, for example, were humourous tweets about a negative situation, thus they were counted twice). This suggests a balance . However, only 3 tweets in the Desktop timeline were truly neutral (neither positive nor negative) and none of the mobile tweets.

***

So QuillConnect wasn’t wrong, but it was reductive. It was limited by lack of detail, which a human user or Twitter’s own algorithm seems able to deliver on. The difference between ‘balanced both positive and negative neutral’ and ‘not expressing positive or negative neutral’, for example, is significant.

QuillConnect is unable to analyse sophisticated or contextual content. It suggests: “For example, the most retweeted of your followers have utilized #westconnex in their recent tweets.” This is because I follow a number of accounts campaigning against the East-West Link (an unpopular, expensive proposed toll road through Melbourne), and WestConnex is a similar road in Sydney. Some of the people I follow (who follow me back) are active campaigners against both roads. Me jumping on that bandwagon to increase my reach would be strange and inappropriate.

Twitter’s Discovery timeline alogorithm, on the other hand, assumes I am more interested in historical artifacts, more likely to click on an article about social justice, or online security, than on a campaign going on in my backyard. They are probably right.

***

So, algorithmically narrative science is still ‘uncanny‘. It is creepy.

The QuillConnect algorithm is still too far off, and is therefore producing negative emotions.

Uncanny Valley, via Wikimedia. See http://www.androidscience.com/theuncannyvalley/proceedings2005/uncannyvalley.html

The ‘uncanny valley’ suggests that a ‘bunraku puppet’ is clearly unhuman but would still garner a ‘positive’ familiarity response.

Discovery on mobile is also somewhat creepy–these would not have been the tweets I picked out for myself as being the ones I’d like most to see. But the Discovery timeline (when it has enough data) is starting to climb out of the uncanny valley.

It is not yet human, I still probably won’t use it, but it didn’t annoy me much and it’s analysis of my interests was ‘close enough’. 

Every semester, there are three weeks that are absolute hell: O-week to the end of week 2. Week 3 is busy, but less stressful. Every year, I make sure I see my friends before O-week starts, and then tell them I’ll see them again in a month.

This blog has had the same problem. It’s not that I haven’t been doing anything, but it’s been out of synch, late, or patchy.

I’ve updated the things I was supposed to have done, or had started but not published, in week 8 only this weekend, as we turn the corner into week 9. I managed a couple of tweets in the tweetorial (and neither of them had space for the hashtag, #fail).

I did have a fascinating conversation with Nicholas and Jeremy earlier:

https://storify.com/katrinafee/foucault-and-algorithms

I had trouble finding an algorithim I could play with, so after a few abortive attempts, I finally worked out I could look at Twitter’s Discovery timeline. I commented on Jin’s and PJ’s bogs.

Posting less often means that some of my posts have been longer. I’ve been continuing to document the places where I study (and increasingly, the food I eat as I study), including my first video:

Anyway, week 3 is coming up. Busy, but no longer crazy. Hopefully that means I’ll have a chance to catch up.

I selected the the top 10 tweets in my Twitter Discovery timeline in both Mobile and Desktop versions, at 1.30pm Sunday 15 March. (They are listed below, under Appendix). I carried out a content analysis of the tweets to see why Twitter might serve them to me.

***

Content analysis: coding

Analysis.

All the tweets included some visual component: a screet shot, picture or gif. Most of the accounts are large, established accounts belonging to publishing platforms or authors.

I did a quick content analysis, where I coded the text, image, embedded content, or information from the link (but not following the link) in each tweet into:

• Geek culture
• Gender or cultural Diversity
• Privacy/online security
• Technology
• Hard News
• Science
• Pop culture
• History
• Writing
• Visual attraction (pictures that accompanied the tweets were pretty, rather than funny or screen texts)
• Food
• Academia
• Humour
• Inspiration

(These categories are in order of identification, not anything more significant).

Tweets were coded with as many of the above catagories as might be relevant. Saladin Ahmed is a fantasy writer and geek/pop culture critic, he writes about racial, gender and religious diversity, and he often posts humourous tweets. He appears at the top of both lists, once more on desktop and twice more on my mobile list. His top tweet registered in 7 of the above catagories (the highest of any tweet).

Twitter thinks this is my ideal Tweet.

There were 5 tweets coded Humour on mobile and 4 on desktop. 4 tweets coded Geek culture (Desktop: 3) and Diversity (Desktop: 2) on my mobile timeline. 3 tweets coded Privacy/online security (Desktop: 2) and History (Desktop: 1). Overall, mobile Discovery tweets registered 34 categories, where as desktop tweets only fitted 28.

Therefore, categories that met my interests were more likely to be served to me in the Discovery timeline of the mobile app than the desktop web version. 

However, this left a large gap–what was influencing the desktop version to rate these particular tweets highly?

I had included the information Twitter shares in grey above a tweet to tell me why it’s in my timeline in the descriptions (See appendix). One clear difference stood out so I added another emergent category:

• High Retweet count

Of the tweets in the desktop Discovery timeline, 5 had a high retweet count (66-197). Of the tweets in the mobile app, there was only 1. This accounted for most of the gap between desktop and mobile.

In the absence of individual engagement data (suggested by what I favourite, RT, click on, reply to etc), Twitter uses general popularity as a category to decide what is going to appear on my Discovery timeline.

And yes, I did make pie for Pi day.

I made pie for 3.1415 day.

***

Appendix 1: The top 10 tweets

1. Saladin Ahmed (who I follow, RTd by another person I follow) on the Telegraph pointing out that now, gasp, white men are being targetted for online abuse too. [Desktop & mobile]

2. Slate (who I follow, RTd by 147 other people) on Pi day 3.1415.  [D]

New York Times (who I follow, RTd by William Gibson) on how the ‘tech titans’ of Silicon Valley who created platforms that require and harvest personal information are protecting their own privacy. [M]

3. Bibliophilia (who I follow, RTd by 66 other people) on an 18th century double bible. [D]

Newsweek (followed by someone I follow) on whether the Antropocene started with the Native American genocides [M]

4. The Economist (who I follow, RTd by 197 other people) on the increased risk of nuclear war.

Classicpics (RTd by someone I follow) on the ‘perfect body’ in 1955 (the woman is a healthy weight, with muscles and curves). [M]

5. New York Times (who I follow, RTd by William Gibson) on how the ‘tech titans’ of Silicon Valley who created platforms that require and harvest personal information are protecting their own privacy. [D]

Slate (who I follow, RTd by 147 other people) on Pi day 3.1415.  [M]

6. Saladin Ahmed (Rtd by 2 people I follow) with a Star Trek reaction gif about Gizmodo, diversity. [D]

Science headline of the week about snail sex (RTd by someone I follow). [M]

7. Science headline of the week about snail sex (RTd by someone I follow). [D]

Saladin Ahmed (Rtd by 2 people I follow) with a Star Trek reaction gif about Gizmodo, diversity. [M]

8. Wall Street Journal (who I follow, RTd by 166 other people) on the return of the business phone call. [D]

Endgadget (followed by someone I follow) on the fact that new HBO still ‘has strings attached’. [M]

9. Advice to Writers (who I follow, RTd by 82 other people) with a quote from Maya Angelou. [D]

Person RTd by someone I follow on 11th century limestone head discovered in Norfolk.[M]

10. An academic I follow (favourited by another academic I follow), joking television marathons should count towards tenure. [D]

Saladin Ahmed, with a reaction gif from Lord of the Rings about having too much work to do. [M]

Gillespie explores in more detail the reason that Knox (2014) is a useful insertion into the field. Algorithms are broadly understood as, and presented as, objective and data driven. However, algorithms are actually crafted by software engineers, and are likely to emphasise their world views as to what is important, and what counts.

Gillespie (nd) p. 2

Gillespie (nd) p10

Gillespie (nd) p. 12

For example, monetisation, popularity, or scientific evidence (see recent changes to Google for health searches), which are Western, capitalist and post-Enlightenment values. Other voices may be hidden because the algorithm or software doesn’t even count it–the new Health app on all iPhones was launched with no way to track menstruation, even though this is a very common aspect of women’s health. This suggests male is another human bias baked into many algorithms. Gillespie mentions other such biases, such as Amazon and YouTube ignoring ‘adult’ or ‘suggestive’ (i.e. with sexual content) works in their rank (p.5-6).

Algorithms are not exhaustive knowledge systems, but fast heuristic devices, where quick, good enough, judgements are preferred. This has the effect of privileging norms and majorities, and therefore increasing their significance.

What is being posted on the web is strongly influenced by the algorithms. I continue to be part of various Social Media / Community Manager online communities. Recently, Facebook started to prioritise ‘native video’ (ie video posted directly to the site, rather than embedded from another site). They are apparently strongly promoting posts with native video, meaning that where 4 years ago your most effective posts had a picture (as they were promoting Facebook as a visual platform) now they will have a video.

Content managers and marketers are therefore out shooting video where we used to go out to shoot pictures. Their daily actions and tasks are changing. This also changes the way we look at the world. I often framed things I saw as I walked around campus as candid phone shots, or later framed and filtered Instagram pictures. Looking for video is a different way of judging what we look at, literally a different way of looking at the world. (See p.20).

***

I made a video with Instagram to reflect the ways in which the above might play out in digital media.