Andreas Weigend | Social Data Revolution | Fall 2014
School of Information | University of California at Berkeley | INFO 290A-03


Audio: weigend_ischool2014_6.mp3 (dropbox)
Transcript: weigend_ischool2014_6.docx (not done yet)

Timeline Oct 28, 2014 (Part 3)

5:40 BEGIN
Guest: Ari Bader-Natal

6:15 If you had all the data...
6:25 Logistics / Housekeeping
6:30 END

7:00 DINNER with students and oDesk founder and Chief Data Scientist at Nefeli Café (1854 Euclid Ave)

New age of Education

Recently, new forms of education are growing rapidly. Especially, MOOCs (Massively Open Online Courses), which are free online courses provided by famous universities are spreading worldwide. For example, the largest MOOCs provider Coursera has more than 10 million users, providing more than 800 courses.
Example of MOOCs (Udacity)

A recent MOOC that UC Berkeley has recently been more involved in is edX:
These online classes allow anyone, whether they are a registered student at another university or not, to sign up and take free courses. There are additional support systems, such as bulletin boards and online "office hours" in which students can interact with each other and the staff. There are graded homeworks and exams with feedback. In addition to making education more accessible to the general public, edX has gained popularity through the schools that use it, which encourage those students to try out new courses online that they do not have to commit to.

As well as MOOCs, there are also some other forms of digital solutions for education, such as digital textbooks and software for learning languages.

Example of Digital Textbook

Example of Software for Language Learning (Rosetta Stone)

These technologies can change education dramatically. First, one is now able to learn anywhere, anytime at a very low cost. Today, a wider range of people such as workers who don't have time to go to school, or children in developing countries who cannot afford to go to universities have access to education too. Also, the combination of multimedia and interactive tools made the learning experience more exciting, and learners are able to learn more deeply and at their own pace.

In addition to the above benefits, digital learning platforms enabled us to collect various data about how learners learn. Among MOOCs providers, edX is particularly interested in this area. edX is a non-profit service and one of its goals is to conduct research on learning science using data obtained from the platform. edX has published many research papers as the outcome.

Minerva talk by Ari Bader-Natal

Minerva tries to reinvent the university experience through a set of advanced new technologies to help capable students access and engage in quality learning. They admitted their first class of students this year aiming to have them go through a full college education using the new Minerva platform and methodology.

Minerva relies heavily in data and their in the business of trying to use that data to improve the way their students learn. Since much of the academic information can be found on the web and many tools are readily available for that purpose -just look at the rise of the MOOC- Minerva then can focus on fostering the intellectual development that occurs on top of the content delivery. This is a new approach and an interesting model that benefits from the MOOC phenomena rather and trying to replace it.

Some characteristics of the Minerva platform:
  • Informal Class structure
  • Classes are between 15 and 19 people
  • Gesture sensitive for students and professors
  • When someone raises their hand (metaphorically) the instructor can also see a one liner of what the person wants to say


Platform design focus:
  • Designed with a focus on intellectual development and individual thinking
  • Aims for students to communicate effectively, think critically and creatively
  • Active learning techniques to get more value out of the class
  • Be able to transfer these teachings to other parts of your life


Metrics that we might look at during their experience with Minerva:
  • CLS assessments tests as benchmarks
  • Measurements on specific skills
  • Highlighting specific moments from the professor


Minerva uses the flipped classroom model. This is a challenge for teachers as well - they need to adjust their methods. As part of the new model teachers need to go back after every class and revise results and critical moments


Questions and Answers

Q: Are you doing any measurements in real classrooms?
A: Not yet.

Q: How is the faculty's difficulty to adapt to the new digital tools?
A: We need adjustments both for faculty and students.

Q: Are you thinking of applying your techonology to corporate training?
A: Currently, we are focusing on universities.

Q: Why universities?
A: Our goal is to rethink higher education.

Q: How do you decide what features to implement?
A: Ask for feedback through the network of faculties.

Opportunities for digital learning
The use of smartphones, tablets, and other tech items in the classroom do not necessarily have to have a negative impact on student achievement. On the contrary, the increasing accessibility and growth of technology presents teachers with the unique opportunity to take advantage of those once distracting gadgets, and use them to facilitate academic achievement in new and innovative ways. It's also a great opportunity to bring data into the classroom. In this capacity, teachers do not need to be constantly fighting for student attention, but can freely accept it, by introducing a new educational environment that will automatically encourage student participation (and therefore produce more datas that would enable the teacher to use new tools).

Below are some resources that we find useful when attempting to implement technology into classrooms. It might be implemented in the next SDR class, or in yours, when you'll be sharing your knowledge.


Socrative Teacher

Educators can initiate formative assessments through quizzes, quick question polls, exit tickets and space races with this app. Socrative will instantly grade, aggregate and provide graphs of results to help the teacher identify opportunities for further instruction. It definitely helps visualize student understanding.



In my opinion, the most powerful tool to build e-learning class supports. It's basically a Keynote specifically dedicated to class. For those of you that are already familiar with coding, that might not be really useful, but if you want to use great tool, without any efforts, that's for you.
I recommend the video on their landing page to have a better idea of what they're doing.

Learning Management System

So this wikispace is great. Because it's well-organized, and also well-maintained by both the teachers and the students in an amazing collaborative effort. Though, they are tools that are even more powerful than that, and that have a better design. Check them out, you will be surprised by how beautiful and joyful learning can be with digital.

My two favorites are Engrade and MyBigCampus, mostly because they emphasizes on the production of data. Not only they're great Learning Management System, enabling a great organization of the class, but they also provide data mining tools to improve our efficiency.

I recommend that you watch their videos to understand why school and learning will never be the same again.

Digital Mentoring

Note: The audio and video of this part of the class can be found under class 7, Fitness

Guest speaker Mike Freed from SRI talks about using a data science approach to creating a digital mentor.

With the emergence of the internet, we have access to an almost infinite amount of learning material, yet the task of learning in many ways hasn’t become any easier. It is not enough that the material is available to us we still need some guidance for how to approach it.

Intelligent tutoring systems

Machine learning is useful for helping us learn facts
  • It can present questions and determine how well we are doing
  • It can calculate what the next question should be
  • It can keep track of progress
Examples of intelligent tutoring systems include:


The problem with the tutoring systems is that often it is not the learning process in itself we have a problem with, but finding the motivation and knowing what to learn. Especially when concerned with hard problems.
  • To be effective a tutoring system needs to affect our behavior
  • People blame lack of time for not learning new things – a tutoring system doesn’t help with that
People apply different strategies when learning
  • Strong self-learners – have intuitive understanding of learning strategies
  • Weak learners – need these strategies to come from somewhere else

The marshmallow experiment tested kids understanding of delayed gratification and studied the strategies applied for overcoming a hard task.
  • These strategies greatly impact how well people are doing in life


To broaden the role of the intelligent tutoring systems we can talk about mentors.

What is a good mentor?

  • Someone who pushes you in the right direction
  • Uses experience to open your eyes
  • Helps the typical learner apply strong-learner strategies

PERLS: Pervasive Learning

At SRI Mike Freed i working on a digital mentoring system called PERLS.
Building an app that function as a personal mentor in the process of onboarding new employees. This process can normally be very stressful and confusing for people, since there are so many new things to familiarize yourself with.


By using an app, you avoid having to constantly disturb your new colleagues with problems that to them seem trivial.
A digital mentor will be non-judgmental and doesn’t have more important things to do.

The app contains several different types of information ”cards”, e.g.:
  • Quick hint
  • Surprise me
  • Quick picks
  • To Do

The digital mentor doesn’t just make the learning material available to you, but it attempts to sell it to you by being highly suggestive and by only recommending things that are of interest to you.
For the recommendation system, it can in theory tap into several data sources like:
  • Calendar
  • Physical location
  • Email exchanges
  • Current project
  • Previous experience
  • Learning process

The learning material is based on routine, but it is part of the vision that it someday could be generated through crowdsourcing.
  • The system know a lot about who is an expert in what, so it can create activities for those people, which implicitly let them generate new material.
The interesting parameters in regards to the equation of learning in this context is
  • Motivation
  • Spacing effect
  • Social pressure
  • etc.

This page initially created by:
Yu Suzuki
Colas Zibaut
Nicolas Soldi