There is a paradox: Humanity’s most developed organizations and systems are based upon what is learned in our education systems; yet, the field of education lags behind nearly all others. One such area I have seen, is how feature-poor Student Information Systems (SIS) are. Despite such systems being case studies in many database books, most of these systems do not use any data science methods to improve operations. Specifically, I have usually not seen active security, predictive analytics, nor even resource optimization as features. Here is why these are important to have, and my invitation for SIS providers to come into the 21st century.
First, before I get started. There has been a woeful state of many Student Information Systems in general. For example, one very popular SIS in California has nearly all of its tables and fields only using 3 letters to designate them, making it very difficult to quickly know the meaning of data. I have seen other school data systems (although not an SIS), not have normalized tables, and up until recently use Fox Pro or Microsoft Access as its back-end. (Thankfully most of these finally changed to having a better back-end, and being better designed). But, just storing data in a manner that can be easily retrieved and used for day to day operations is not enough. If we want education to improve, we need to start to use data better. And while all SIS providers are saying they are doing this. They are usually doing it in very rudimentary ways, and not utilizing methods that sometimes have been around for years.
What I’m about to talk about is actually a form of predictive analytics, but it is so important and so lacking that it deserves to have its own title and explanation. With so many hackings that are occurring every day, the state of security in student information systems is abysmal. While most have implemented known “best practices”, such as encrypting passwords (although I know of a very popular SIS for California charter schools that doesn’t do this!), I have not seen an SIS yet that monitors its data traffic, looking for signs of illegitimate use. Rudimentary forms of this could be easily implemented by looking for attempts to snarf the whole database, as most hackers would rather grab all the goods and go, so they have less chance of getting caught, than to stay logged in for longer periods of time. More advanced features could be used similar to how banks look for suspicious credit card activity etc. The fact that this is a rare or possibly non-existent feature is very bad, given that nearly every person in the United States has some connection to a school’s SIS, where their data is stored because they are a parent and/or student (whether or not they have any current enrollments, as most SIS have eternal archives, so that transcripts can be produced). While many may think school data is safe because there have been relatively few hackings in the news, I think the opposite may be the case, that there may be far more hackings that have occurred, but the security is so bad, that often these hackings are not even known.
Predictive Analytics for Educational Improvement
While some eLearning platforms are starting to use data science methods to determine what students know and don’t know and trying to have on-demand and one-on-one learning environments, there is a general lack of using these methods in the student information system itself. One area that could crucially help many high schools and colleges (as well as other postsecondary schools) is to utilize predictive analytics to find students that are at the most risk of dropping out based upon their behavior. Similar to how predictive analytics work in other contexts (with spam filtering being a common one that most people are familiar with), Bayesian methods could be used to look at the probability of each individual student dropping out, based upon known data. The priors (initial probability estimates) could be based upon correlations found in scientific studies about drop outs.
Using such a system, when a school found sufficiently high probability that a student may drop out they could take appropriate interventions. This could save the student from having a high probability of having a life of poverty, and at the same time save the school tens of thousands of dollars of lost revenue.
Although, it still amazes me that given that there is a co-aligned moral and financial benefit to preventing students from dropping out, that more schools and districts don’t yet clamor for this technology. I think sadly, most administrators can’t even conceptualize this yet. But as Apple Computers has shown over and over again, people often don’t know what they want… yet. So I believe the SIS that implements this, could be the “killer app” for education. (Although, the cost of switching SIS providers is so high, because of training, knowledge level, transitions of records, that it is hard to say whether existing clients would switch unless the SIS could show how this feature had a far better cost-benefit ratio)
Optimization of Resources
This is one area where some Student Information Systems often partially implement; generally in the area of scheduling students. But, given that linear programming has been around for over 75 years, it would seem that the ability to optimize scheduling based upon constraints of prerequisites, classroom size limitations, credits needed for graduation, and what classes teachers could potentially teach, should be far better and fully implemented than what it often is.