Predictive, Descriptive and Prescriptive Analytics and the Movies


Desk Set still



Tracy, Hepburn and EMERAC in DESK SET, 1957

I watched the 1959 film Desk Set over the holiday break. It is set within the TV network FBN, Federal Broadcasting Network (the exterior shots were done at Rockefeller Center, headquarters of NBC). Bunny Watson (Katharine Hepburn) is in charge of its reference library, which is responsible for researching and answering questions on almost any topic. With a secret merger pending, and anticipating a lot more demand for the department, the network boss has ordered two new computers.



Of course, this being 1959, the computers are called "electronic brains" in the film and they are huge. Richard Sumner (Spencer Tracy) is the inventor of them and they are called EMERAC. That name is some wordplay from ENIAC - the Electronic Numerical Integrator And Computer that was the first electronic general-purpose computer.



I also saw a new film over the break - The Imitation Game based on the book Alan Turing: The Enigma. The ENIAC computer was considered to be "Turing-complete" - a term from the work of Alan Turing. In the book and film, set during WWII, Turing is trying to crack the German Enigma code and in the course of doing that, saves the Allies from the Nazis, and sort of invents the computer and artificial intelligence.



The Spencer Tracy character in the 1959 film was also trying to create a digital way of solving problems. They describe him as being an "efficiency expert" which was a new and big concern in the 1950s.



Today, predictive analytics has become a big topic in educational technology and I have written a number of posts about its use in education. It is a way of using statistics, modeling and data mining to analyze current and historical facts in order to make predictions about future events. An example of one of the desired educational uses is to monitor at-risk students and allow interventions at the proper times.



 



Data analytics in higher education is still in its early years and the terms have changed over the past few years as the use of the term "big data" has replaced "data mining" in popular conversations. Where I was once reading articles about using "descriptive analytics" - the analysis of historical data to understand what has happened in the past - now I'm more likely to find articles on "predictive analytics" - using historical data to develop models for helping to predict the future.



Prescriptive analytics takes those predictions and goes to the next step of prescribing recommendations or actions to influence what happens in the future.



Confused? As an example, using big data and descriptive analytics about students and any particular student, we might predict the student's performance and problems in the current semester and then using a prescriptive analytics-driven learning management system we could recommend additional material, resources online or even notify on-campus people and departments to interact with the student early on.



Prescriptive analytics seems best-suited for educational problems like student retention, enrollment management, prospect analysis, improving learning outcomes and curricular planning. These are all problems that can be addressed with data analytics because there is adequate high-quality data to analyze those problem.



Did you read Moneyball or see the film version of Michael Lewis' popular book? It can be viewed as the story of the power of predictive analytics as he describes baseball's Oakland A’s team manager working with the lowest team budget in Major League Baseball and using predictive analytics techniques to turn around his team’s performance.



What schools need to do is similar to the "business rules" that companies formulate with input from various stakeholders in the organization.



Sometimes the data may produce results that are open to interpretation and campus experts need to be involved. In an article, "Prescriptive Analytics for Student Success," it is pointed out that student data now includes that generated by mobile device usage, campus cards, social media and sensor technologies. It presents an interesting case of an on-campus student who is not using dining services as often as before. Does that mean the student isn't as active socially on campus? Is she depressed? Can we add to the data class attendance or even clicker use? Is she more likely to drop out?



These "alternative data sources" are still emerging and may cross over into FERPA and privacy concerns about what is permissible in data collection. 



Learning analytics is another term used and seems to apply more to using learner-produced data and analysis models to discover information and social connections for predicting and advising people's learning. This area sounds like it would appeal more to teachers than the earlier examples, but in some ways there is a lot of crossover in studying individual learners. The data might allow the learner to reflect on their achievements and patterns of behavior in relation to their peers. It can warn them of topics or courses requiring extra support and attention. It can help teachers and support staff plan interventions with individuals and groups.



For departments and institutions, it can help improve current courses, help develop new offerings and develop marketing and recruitment strategies.



Predictive analytics is a big field and one that seems to strike fear into teachers much like those computers in Desk Set did more than 50 years ago. It encompasses statistical techniques from modeling, machine learning, and data mining. Do we trust the predictions about future, or otherwise unknown, events?



This is being done in actuarial science, marketing, financial services, insurance, telecommunications, retail, travel, healthcare, pharmaceuticals and other fields, but many educators (and I confess to still being one) are hesitant about moving business models into education.



You may be able to accurately do credit scoring (as in the FICO score) in order to rank-order individuals by their likelihood of making future credit payments on time, but can we really predict how a student will do next semester in Biology 102?



Nate Silver gained a lot of attention with his blog and books, especially during the last presidential election, showing that predictive analytics pays big dividends in politics, sports and business.



Companies such as Google, Twitter and Netflix are hiring predictive analytics professionals to mine consumer behavior and they are in the front of the office rather than crunching numbers behind the scenes.



In higher education, student retention is such a big concern that colleges find success using data analytics, it will quickly find its way from administrative tasks and into classrooms.



 


Trackbacks

Trackback specific URI for this entry

Comments

Display comments as Linear | Threaded

No comments

Add Comment

Enclosing asterisks marks text as bold (*word*), underscore are made via _word_.
Standard emoticons like :-) and ;-) are converted to images.
BBCode format allowed
E-Mail addresses will not be displayed and will only be used for E-Mail notifications.
To leave a comment you must approve it via e-mail, which will be sent to your address after submission.

To prevent automated Bots from commentspamming, please enter the string you see in the image below in the appropriate input box. Your comment will only be submitted if the strings match. Please ensure that your browser supports and accepts cookies, or your comment cannot be verified correctly.
CAPTCHA