Profile cover photo
Profile photo
Alfred Essa
320 followers -
Vice President, Analytics and R&D
Vice President, Analytics and R&D

320 followers
About
Alfred's posts

Post has attachment
Alfred Essa commented on a post on Blogger.
As a former CIO, I am surprised and not surprised that this (ERP botched implementations and cost overruns) is still happening. Also, annual operating budgets for ERP (staff, licenses, etc) dwarfs by at least one order of magnitude budgets for learning systems (e.g. LMS) and instructional materials. Yet everyone complains about the high costs of learning systems, textbooks, etc. When I started my career as a CIO at a small college, my budget for academic computing was $10,000 and for administrative computing it was $500,000. It seem not much has changed. And who is held accountable for these catastrophes?

If one analyzes IT budgets in higher education I will wager that the portion that goes to support "administrative systems"is a giant sucking sound. It is also one of the reasons, not the only reason, why higher education costs are out of control.

eLiterate can shed some light on this by doing a neutral budget data analysis. I am willing to lend hand with the analysis,. But I am biased: I work for a vendor.

I would note that it's easy to blame the CIO and offer her as the scapegoat. If one digs deeper, successful implementations are driven and led by the business units (registrar, admissions, financial aid, HR, finance etc.). If they don't have their act together and not leading the implementation, it will inevitably fail. ERP implementations are not "IT projects" they are strategic projects that have to be led by the business.

One of the dirty secrets of higher education then: cost of administrative computing or ERP.is a giant sucking sound. Also, whatever happened to open source (e.g. Kuali). Wasn't is supposed to solve this problem?

Sorry for the rant. My views are my own and not my employer

Post has attachment
Alfred Essa commented on a post on Blogger.
Michael,

You know my personal view on software patents: I strongly oppose the offensive use of patents and condemn it.

I can't speak on behalf of McGraw-Hill Education, but as a learning science company we are increasingly moving towards openness. Closing knowledge is antithetical to science. Deterring others from building on your knowledge is also antithetical to science.

We want our customers and other researchers to know how our products work. The algorithm ("Knowledge Space Theory") behind our ALEKS adaptive suite of products, for example, is open. Researchers have continued to work on it and published their results in peer-reviewed articles. Our researchers at MHE are expected to publish and share their claims and results with other researchers and the learning science community.

I agree completely with Mike Sharkey that it's execution, implementation, design, user experience, and support that will differentiate our products and solutions. Tens of thousands of engineering and design hours have gone into ALEKS alone. That talented work can't be easily reproduced. And if our competitors catch up, we will go faster. That's healthy. It will also accelerate scientific discovery.

I will also be pushing the community for a much higher bar when it comes to openness. When one makes a scientific claim, it's not good enough, in my opinion, these days to have published your result in a peer-reviewed journal. Your data, your computation, and your code should be available to other researchers for review. Anything less is not science,

Post has attachment

Post has attachment

Post has attachment
Alfred Essa commented on a post on Blogger.
Phil,

I wonder if you are not throwing the baby out with the bathwater. There are two distinct questions. First, is there market disruption and can new technologies catalyze them? By market disruption we mean entire industries can change essentially overnight. There are many indisputable examples of this, or at least I think so. Amazon disrupted and decimated traditional book stores. Uber has disrupted the taxi industry. The PC disrupted typewriters. Emails disrupted the postal and greeting card industries. Second, we might ask how disruptions occur. This is where Christensen comes in. He proposed a model of how disruptions occur. I think most people who study market disruptions for a living don't believe that Christensen's model is the only one. Also, we might believe in Christensen's model but not believe that it explains all disruptions. In my opinion, Christensen did have a salutary effect, namely warning incumbents that you could be innovating (in one sense of innovation) and still get destroyed. 

By throwing the baby out with the bathwater I mean this: if you want to throw out Christensen's theory of how disruptions occur, then go ahead. He also believes that technology will be the primary source of disruption in education. That's also arguable. The baby is the belief that the education industry is prime for disruption.  

I have read Audrey Watters post several times and I am not sure what she is arguing. At times she seems to be arguing against Christensen's model. At other times she seems to suggest, but I don't know whether I am misreading her, that disruptions don't occur?

Al

Post has attachment
Alfred Essa commented on a post on Blogger.
These types of claims make me cry statistical tears. SOB.

Post has attachment
Alfred Essa commented on a post on Blogger.
Although statements by their spokesmen are at times confusing Unizin's goals seem, at least to me, very clear.

What is Unizin? "A cloud-scale services operator for content, platforms and analytics steered by its members." 

Unizin has already standardized on Canvas at the platform layer. The content and learning analytics layer are next. According to their timelines a "rudimentary repository" will be ready in Spring 2015 and a "rudimentary analytics platform" will be available in Summer 2015. Of course, the three major components will inter-operate using open standards. Unizin's endgame, therefore, doesn't seem to be a mystery.

What's not clear is when the "cloud-scale services" will come online, how the three component layers will be architected (e.g. what does "loosely coupled" mean exactly?),  and when member institutions will start using it in production. 

Post has attachment

Post has attachment
New Data Analysis Tutorial: Descriptive Statistics in Python / Pandas
Wait while more posts are being loaded