How will the field of education be affected by data science?

Data science as we define it 'captures structure or unstructured data' often through a scientific method; whether it is qualitative (opinions or perspectives) or quantitative (numerical values that reflect points in or over time). Currently, education is affected or influenced by data science in a few major ways. In a brief, here are a few examples.

Technology- A LMS system can track, monitor and report large amounts of user data for instructional analysis and feedback.

Marketing- Quantitative or qualitative data is used as evidence to persuade people either for the purpose of education sales or recruitment for participation of education initiatives that affect student academic outcomes.

Trend Insights- Different types of data are often used to determine strategic planning for small or large projects that help with hiring new educational staff, faculty or reform efforts in the area of diversity and inclusion.

Adaptive Technology- Data science has been a huge help to companies like DreamBox, that rely on data to assist with modifying lesson plans (adaptive technology) based on a student's learning progress. For example, as a student moves through an adaptive software lesson plan, the lesson plan will not advance to new problems unless the student demonstrates competency in the initial problems presented to them.

Hope this helps!

Increased data science in the field of education will produce greater evidence of effective learning practices and technologies for implementation in the field. In my experience, greater data collected through research and evaluation has proven to improve the quality of instruction and learner outcomes.

There is an entire emerging field of learning analytics dedicated to analyzing data scraped from learning management systems and figuring out what works best for different subjects and types of learners. In the (not too distant) future, there could be an assessment about how you learn best that you take at the beginning of a course, and then your content, activities, and assessments would be curated differently based upon your learning preferences.

Another way education will be affected by data science is the influx of linked data about jobs. As education and the workforce continue to blur lines and converge in the future, it will be important that learning outcomes are connected to the changing needs of the workplace. The current standards being used (O*Net) to align learning outcomes to workplace competencies are only being updated about every 6 years. As we know, in the future of work, technologies will be changing at an exponential rate, impacting the competencies needed. There is a need for realtime data that reports back to educational institutions what skills need to be taught.

As you can probably guess, LMS software is becoming more and more capable of tailoring a virtual learning experience to the learner based on their performance. Not only can the software run reports for us to know where all our learners are in a program/how they performed, but we can also program these systems to provide extra courses, tailored additional feedback, or customized learning experiences based on the input they give during a program.

Another thing that's changing is the expectation of measuring learning results. It used to almost be above and beyond if we could display exactly what behaviors changed as a result of learning. Now, with analytics as powerful as it is, it's a given that you will be expected to collect data that demonstrates that your training actually produced an outcome (worked).

Everyone else has to measure job performance and provide clear performance data for their employees, too. That means when you work for a client or a company to design training, they will often have data that empowers you to pinpoint where the weak or blind spots are for your audience, and design specifically to those learning objectives. It puts a lot of power in your hands to offer targeted training. For instance, being asked to "fix the sales team" is much more vague than if they say "here are our key performance metrics from the last 12 months, and where we saw the gaps. Can you teach to these gaps?"

Data is generally our best friend in this industry. I think the rest of the workplace is finally catching up in a way that lets us do our job in a spiffier, more awesome way.

I'd strongly suggest you investigate some of the free content on https://www.elearningguild.com/. They are the industry leaders on researching how tech and learning intersect. DEVLEARN, their annual conference in Las Vegas, would be an outstanding opportunity if you have an employer who will send you.

If you are interested in entering the field of education from a data science point of view the best place to start is to research and track the evolution of the use of xAPI tracking. In simple terms, this is the way that a person's learning habits are tracked in an LMS and it can also be used in developing 'optimized' learning paths for a learner, sometimes known as personalized learning.

The key to understand is that xAPI allows educators to track what is called 'informal learning', that is, what is accessed outside of a formal course/quiz sequence. Prior to the development of xAPI the SCORM standard only could track those formal courses. Now we can track any 'optional resources' such as a video etc, that is accessed outside the course. What this means is that we can, in the long run, track how learners with a high performance rate learn. This will then influence how we teach new learners.

So you can see, data is used not just to 'gate' a learners path such that they only access the next level of learning when they 'pass' the current level. That has been happening for awhile.

The next best thing that big data will offer us is to learn from the learner 'best' education or training practices.

One other use of AI/ML and big data is to create 'automated' learning that combines both of the former. It learned from all learners and combines that modeling with learning from the current learner and personalizing the learning from there. There is a cool diagram and explanation of an example here: https://www.dev4x.com/ourapproach#moonshot-education-project-detail

Anyway - this is what I am tracking. Feel free to reach out to ask more.

Data science will be one of several new competencies across most sciences. I recently wrapped up a specialization in data science on Coursera and the topical coverage was vast, from machine learning to statistical inference to R programming. You see this happening today throughout the K-12 curricula, and pointedly so in most lower division college courses. Many college breadth requirements now include a data science course or at least a course that exposes students to ds.

I agree with others about how our research methods and evaluations can be enhanced with data science. But let's not forget that a strong understanding of data science presents opportunities where they were once to cumbersome to collect and tabulate.

I feel that data science will only enhance this field. I am now using VR and AR in my design, and this captures the attention of the students. Every year technology increases and gives an ID more to work with. I don’t see how Data Science will replace this.

Data is everywhere in the field right now! I'll share a few examples that can hopefully help you see the scope of how important understanding data science is to education right now.

I'm an instructional designer at a software company, so a lot of the learning content I'm creating goes straight to our customers. There are so many different types of learning being requested of us that we tend to use data to decide what projects to prioritize. We look at views on our website, trending topics on sites like Linkedin and Google, as well as our company's sales numbers to make sure that when we spend time making that amazing learning content, there is a large audience out there that wants to see it.

Obviously, we don't always get it right. So, we also have data attached to every piece of content we create! This helps us see what types of content are preforming well and figure out how we can make more of those pieces in the future. This data also helps us go to our bosses and say "Hey we had "x" amount of views on our videos last month, so we need more money to buy better equipment and hire more people so we can make more videos!"

These are the two biggest influences on my day to day as an Instructional Designer, and I think they illustrate well how important data, and data science understanding, is for learning professionals.

In technology and marketing as well as production

Data science is incredibly important as a measuring and strategic tool. But a small word of caution to go with the rest of this excellent advice: data still requires interpretation. Any given result is not "good" or "bad" by definition. You still have to do the hard work of determining the relevance of your data, even before considering its actionability. You have to look for complications, problems, possible errors, misuses, etc.

For example, a colleague of mine was recently asserting that our data told us that company associates weren't completing their electronic training. But I decided to dig in to that data and found a bug that was preventing our LMS from recognizing when an associate had completed the course. So we actually had a much higher percentage of employee participation than the data showed. Our data was wrong, and so was any interpretation that had been made as a result of that flawed data. We had to "start over" in our analysis.

All data collection and use has a constant weak link: us. As educators, evaluators, and data scientists, we design the collection methods as well as the interpretation of said collection. We often have a vested stake in the results turning out one way or another. We are often unaware of our biases, of our continuous influences on the whole process. Things like Heisenberg's Uncertainty Principle (also known as the "Observer Effect"), filtering the Signal from the Noise, and a lack of skilled data analysts are among the wide variety of problems for data scientists to monitor and account for. Even if you know enough to understand and validly interpret the significance of a result, you still must be capable of communicating that result to the people who need to understand it (and most likely won't without a compelling explanation at a fifth grade level).

It's your job to be aware of these weaknesses at every step, to try to account for it in your measuring, collecting, and interpretation practices. The people who don't aren't just unproductive: they are actively counter-productive. If you aren't aware of this weakness, you will be incredibly confident of the wrong things; and that's the kind of situation that can destroy jobs, departments, even companies.