Course:ETEC522/sept2011/socialanalytics

Home Welcome to Social Analytics. Presented for ETEC 522 Week 12

Launchpad

Have you ever wondered how companies such as Amazon and social networking/media sites have been able to suggest books and/or products we may like to purchase? Are marketing companies randomly targeting persons? How do they know who to call, or send mailouts to?

And are you wondering how this can be applied to... education?

Welcome to the internet and mobile era of Social Analytics. Before the invention of web technologies, businesses learned about the market from offline data, such as using coupons and point systems. As more people search for what they want to buy online, web analytics was developed to observe our behaviour by tracking web traffic. With the development of social media tools, companies have yet another way of marketing their products and tracking our preferences. Could they also be using it in product design and development?

Amazon recommending books to look at, or supermarkets printing personalized coupons for you based on your prior purchase history are examples of how the internet helps companise to learn about our preferences from our online behaviour. Even the latest Facebook update seems to highlight 'top stories' in your social network, and hides others. While some might find it convenient that information presented to us is tailored to our likings, others might no like that we are living in our own “internet bubbles”.

Whether we like it or not, we are being watched. In this week’s presentation, we will be exploring Social Analytics and its counterpart in the education sector, Learning Analytics. We will examine their attributes, opportunities and stakeholder needs with a view to analysing and designing a few ventures of our own.

We will first explore the tools and the range of sites that are keeping track of our online behaviour. Then, we will explore what is social analytics and how it can be applied to education, as well as existing social analytic ventures. At the end, we will reflect on possible ventures from the point of view of an educator/instructional designer, a student, and an investor/entrepreneur.

What is Social Analytics?

Social Analytics refers to the collection and analysis of statistical, digital data on how users interface with an organization, particularly online.

Over the last decade, social analytics has become a primary form of business intelligence, used to identify, predict, and respond to consumer behaviour. Throughout our everyday lives, when browsing on an online store, using a member card to buy groceries on sale, or sharing special offers from our favourite coffee shop on our social networks, each of us continually drops pieces of intelligence. With nearly every click we make, data about our online activity is being collected; it would be difficult to find a website that didn’t monitor and analyse its usage in some way. Some websites use only one social analytics tool (e.g. UBC uses Google Analytics), while others use many more. Indeed, this site is being analysed using Google Analytics. Social analytics programs enable analysts to glimpse meaningful trends in this mass of data.

For ETEC 522, we are concerning ourselves with two forms of social analytics that have applicability to learning technologies: web analytics and social media analytics.

Web Analytics

In its most elemental form, website administrators use a social analytics service, such as Google Analytics - the most widely used social analytics tool globally - to capture and analyse data including:

   site visits and unique site visits (i.e. unique, independent visitors as opposed to one visitor visiting a site multiple times)
   the pages that are the most and the least viewed
   search terms used to find the site
   physical location of site visitors (city/country) and the time of day that most visitors access the site
   the last page site visitors access before leaving 
   the web browsers and operating systems that visitors use (for instance, the Google Analytics on this page reveals that our team includes three Windows and one Mac users, and two Firefox and two Chrome users!)

This information can be used to identify which parts of a website are effectively serving the site owner’s objectives (“Which links are directing lots of traffic to my site? Should I deepen my relationship with that organization?”), and which are detracting from those objectives (“Why do people leave my site from that page more than from any others?”).

Social Media Analytics

Google Analytics first launched in 2005. In the last two to three years, we have witnessed the emergence of more sophisticated social analytics tools that measure an organization’s 'influence' over social media. These analytics perform this task by collecting and analysing data related to a given organization across various social media sites (“Do people tweet favourably about my company? Do they tweet about it at all?” "What are the major trends in my field that we can see over social media? How can I capitalize on them?"). This data can help to provide useful demographic information on who an organization's audience is. Consider, for instance, the following graphic developed by Viralheat which compares television ratings with the number of mentions a show gets over social media. How do you think the intelligence gathered by social media might be useful to the show's producers and advertisers?

Social media analytics helps organizations to identify which social media tools and strategies are measurably benefiting their objectives -- and which have a neutral effect or may even be hindering those objectives. This data helps organizations measure the return on investment (ROI) of their social media strategies, and to continually plan how to best use social media to their advantage.

The social analytics tools we have thusfar considered have targeted organizations as their primary market; some, such as Hunch (founded 2009) are being developed as a consumer service. Hunch combines social media analytics' canvassing of data over multiple social networks with traditional web analytics' application to predicting what consumers might like to buy based on prior purchases. This advertisement from Hunch also nicely illustrates what Social Analytics can do.

Please take a few minutes to go over our first two activities for the week. It should take only a few minutes.

In the first activity, download Ghostery ( http://www.ghostery.com/) onto your most used browser. Over the course of one day of online activity, notice the pop-up box on most pages that indicate which service is collecting information about your online activity. Then, please complete the class survey on how often your information is being monitored.

The Survey is found here: https://docs.google.com/spreadsheet/viewform?formkey=dHMtREtaRFY1TE85UUJVY3lWTGpRWnc6MQ

The Limitations of Social Analytics

Case study: Klout

While social analytics usefully helps to identify broad trends in quantitative, digital data - and may be used to successfully predict individual consumer behaviour - like all statistics, it provides neither understandings of content (i.e. the content of a tweet), nor interpretations or explanations of behaviour. As with all statistics programs, social analytics requires skillful analysts to statistically test findings to determine their significance, and to offer meaningful interpretations and explanations of them. The ongoing controversy surrounding the social analytics service Klout illustrates some of the limitations of social analytics. It also raises questions about and the proper - and improper - ethical and scientific use of statistical social analytics data.

Klout is a social analytics service that sets out to measure an individual’s “influence based on your ability to drive action” ( www.Klout.com); this metric is based entirely on that person’s presence on social media websites. It assigns individuals a numerical influence score, visible to all members of Klout. Some report that Klout scores have factored into hiring decisions.

(source: socialtechnologyreview)

Unlike many social media websites that one needs to actively sign on to, Klout created profiles for people who happen to be connected to Klout members on other social networks (e.g. a Klout member’s Facebook friends). This included minors. Klout’s creation of profiles for people, especially minors, without their knowledge initiated a broad controversy about online privacy and ethics; Klout has since ceased to create profiles for people without their knowledge. Online discussions concerning Klout have also foregrounded that statistical analyses and algorithms cannot measure intangible, abstract qualities, such as “influence.”

Learning Analytics What is learning analytics?

Learning Analytics refers to the specific adaptation of social analytics tools to enhance teaching and learning.

Learning analytics is a emergent area of interest in educational technology ventures. The Horizon report (2011) forecasts learning analytics as a long-term trend that will be prevalent in education in 4-5 years. EDUCAUSE and the Bill and Melinda Gates Foundation have targeted learning analytics as one of 5 categories for funding initiatives under their $20 million (USD) Next Generation Learning Initiative. A primary objective for EDUCAUSE and the Gates Foundation in promoting learning analytics is its potential to increase high school completion rates (Kolowich 2010).

Online learners continually leave behind digital footprints in their virtual learning environments - particularly with the increasing adoption of course/learning management systems (LMS). At its most basic level, learning analytics involves using a web analytics program, such as Google Analytics, to track students’ usage of their LMS and other digital learning objects, as one way to gauge learner engagement. These analytics are built into some Learning management systems including Vista, used at UBC. Educators can use this data to:

   help them make realtime decisions on how they might modify their course to better suit learners.
   Identify potential ‘at-risk’ students who may need an intervention in order to avoid failing a course module or an entire course.

At this macro level, administrators at school and district levels use learning analytics to gauge students’ performance, and to compare how schools are performing vis a vis each other. These learning analytics programs, such as Almalogic (a Vancouver-based venture that UBC works with) and Global Scholar's Pinnacle Insight, also help administrators to plan interventions if a school is perceived to “underperform” in one of more areas. This is especially relevant - and controversial - in the US in the wake of the passage the No Child Left Behind act (2002), which measures schools’ overall performance through students’ performance on state standardized tests.

A core concern about learning analytics is whether using learning analytics as a primary measure of a school’s performance focuses too heavily on quantifiable performance indicators - particularly test scores - at the expense of other forms of performance assessment.

Just as newer social media analytics programs gauge an organization’s activities and influence within social media, learning analytics is increasingly turning to the quantitative analysis of social networks in digital learning environments. SNAPP (Social Networks Advancing Pedagogical Practice) is one university-based learning analytics program (developed at the University of Wollongong in Australia) that analyses the social networks that form within learning management systems. SNAPP records statistics on not only which students participate on LMS’, and how frequently, but also pays close attention to which students respond to which students’ comments and posts, emerging leaders, whose posts are frequent and elicit much discussion, and outliers, who contribute little. Snapp also provides visualizations of these social networks to instructors and course administrators.

http://research.uow.edu.au/learningnetworks/seeing/snapp/index.html

Learning Analytics and Personalized Learning Dawson et al (2010) note that one pressing issue in learning analytics is that the data mine created is seldom accessed by teachers (or learners, for that matter); it is primarily used at a macro, or administrative, level. Just as businesses use social analytics to suggest personalized options to consumers based on their buying habits, individual teachers can use learning analytics to help them develop adaptive, personalized learning plans for individual students.

http://www.slideshare.net/gsiemens/learning-analytics-educause

The School of One program, in use in several New York City public schools, uses learning analytics to develop personalized mathematics learning programs. The School of One’s learning algorithm conducts everyday assessments of a student’s learning style and math skills, and uses this to produce a personalized learning “playlist” for each student. This playlist is comprised of individual lessons in math, which are put into the order that the algorithm determines is optimal for the student’s math skills development. Certainly, School of One is quick to note that this is intended to supplement, not to replace, an individual teacher’s expertise.

Similarly to how Hunch focuses on social media analytics as a consumer service, other learning analytics programs are being developed for use by both learners and teachers to enhance individual learning. Socrato is one such program that is primarily intended for students preparing to write standardized multiple-choice tests. It performs learner analytics on students’ online practice tests to identify which areas students most need to improve upon.