We sit on the cusp of a new field of research, where aggregating old knowledge is helping us learn new things. What is meta-data analysis, and what does it mean for the future?
The information age has flooded us with a sea of data.
One could suppose it’s a natural progression: as we improve how we communicate we also increase our rate of communication. This translates into more academic, entertainment, and political media content for the average person to consume.
But the massive amounts of data that we amass these days are only one side of a shiny new coin. The other side is what we do with that data, and while it is an easy decision to store data, how we use it is a much more complicated affair.
Take modern AI technology, for example. The advances in deep learning neural networks have been astounding in the last few years, and in many ways those advances have been necessitated by creating software that can gather more data than any one person can analyze. Computers aren’t necessarily overall better at analyzing data just yet, we’re working on that, but they are certainly faster and have a higher capacity for dealing with data than a human being could ever have.
So if we can make a machine that can gather so much data, then doesn’t it logically follow that we can create a machine to analyze that data? Well, some people have, and as the field of meta-data analysis grows, more people are seeing how useful it can be.
Meta-Data Analysis for Dummies
To be fair, some enterprising companies and organizations have already started making meta-data analysis tools. If you’ve been coming here for the past few months, you may have seen our articles on the company Palantir, and some recent news is adding another organization to the list of developers of meta-data analysis tools: MIT.
At the MIT Sloan School of Management, researchers are using new methods to aggregate the results of decades worth of research papers, allowing them to gain new insights from old studies.
Furthermore, the method uses studies from across various disciplines, allowing researchers to reach more well-informed analyses for any given topic. According to Prof. Hazhir Rahmandad, “The value of a broader method for quantitative aggregation of prior research can be immense across many disciplines.” Translation: More data in the aggregate equals a higher quality result.
So, What Have we Learned?
For the purpose of the study, the research team used papers on various subjects such as marketing, energy, and health, but the best results seem to have come from their analysis of the basal metabolic rate (BMR) of human beings.Basal metabolic rate = how many calories you burn not getting out of bed.Click To Tweet
A person’s BMR is the result of numerous factors, which made it a prime target for the research team. Using their new method, the team was able to create an equation that outperforms previous models in estimating a person’s BMR, turning the team’s initial theories on meta-data analysis into relevant empirical truth.
While Rahmandad is optimistic about the future of meta-data analyses, his research is far from over. “For problems that don’t yet have a clean answer, the potential for this model is major,” says Rahmandad, and if the results of the study are any indication, he may be right.
If meta-data analysis gives us the ability to glean new insight from old studies, then the potential for modern research to improve upon the past is higher than it has ever been, making the future seem brighter than ever before.
What new wonders will we glean from old studies? It’s hard to tell, since the potential for meta-data analysis is virtually limitless.