DS Reflection #5

The new paradigm age driven by data science has shown us thus far that big data will continue to advance our understanding of complex systems (adaptive/economic/social) when utilizing scale/measurements and a conceptual framework/model that can guide desired actions.

Aral claims that revolutions in science are often preceded by revolutions in measurement. This is true to an extent. In an article, “Big data, New Epistemologies and Paradigm Shifts”, Rob Kitchen discusses how as we collect more data and more information, an accurate model becomes harder to attain. Data might infer a correlation but not direct causality. To resolve this problem, the way in which measurements are conducted must change in a way that allows data and a model to work together. A new system of measurement will alter new data analytics to fit the desired role. This 4th paradigm revolves around data centered exploration using approaches in empiricism or data driven science (inductive, deductive, and abductive methods/measurements). Thus, it’s fair to say a revolution in measurement often follows a revolution in science. And this revamped system in measurement often helps guide data with models to help better understand complex systems. Adding onto this point Geoff West raised similar arguments in his book, “Scale: Thue Universal laws of Life and Death in Organisms, Cities and Companies.” Geoff West claims measurement has played a central role in the development of our understanding of the entire universe. Data provides the basis for constructing, testing, and refining our theories and models indicate what problem or question the data is trying to tackle. For example, the large hadron collider, a particle accelerator, was developed in an attempt to find the Higgs Particle. The large machine extracted 150 exabytes of data per day. Tring to find the particle in .00001 percent of that mess sounds impossible. But using a model/designed framework to guide the machine over the data, they were able to successfully discover the Higgs Particle. West then reiterates the idea that models are not dead, and only by using models with our data will we be able to make these scientific achievements and better understand the world around us.

Barder claims that economic systems are evolutionary systems. Barder noted that the last 50 years have seen economic success and no economic model has been able to accurately describe or predict it. Take the neon capital growth model (discussed in his paper “Development and Complexity”) for example. It offered technical change on top of capital and labor as a means to model output and it used a conventional economic model as a basis. The technical change had no result or answer tho. Barder describes this scenario like the economy was a dish and had some missing ingredient that all models just couldn’t quite catch. It could be capital, institutions, technology, or different politics maybe. Nobody knew. The economic system as a whole is diverse, complex and overall hard to break down. Many economists argued these models describe the potential of institutions and firms, thus that was why there were no answers or ambiguity in their results. Knowing how complex economic systems are and how we model them in such manners, barder made this claim that they indeed are evolutionary systems. But modeling economic systems is not impossible. In Geoff West’s book on scaling, he describes the use of scaling in data science to model companies, infrastructure/social factors within cities, and even living things. More recently, data used to model cities has given us significant gains in our ability to describe, analyze and predict human development processes. We have learned that less infrastructure is needed per capita in cities as the population increases. This finding is important not only in helping us better understand human behavior but also fiscally as well. Using sublinear scaling for example has allowed scientists to make an educated guess on the characteristics of large companies based on those of smaller ones. Or inversely, companies that make trillions of dollars in revenue, like Walmart, can be scaled down to measure characteristics of smaller companies with under 10 million in revenue. Of course this is just one realm of the economic system but data science has made major major advancements at understanding many of its underlying principles through scaling.

The only main reason to be concerned for the use of data science is most likely data privacy. Measuring phone data and where individuals shop, read information, get the news is valuable information for companies or the government to better understand the population. But it does raise some concern about how much others know. In terms of struggles, data science is still a relatively new field, so there are a couple. There are countless ways data science can be applied but finding a model to narrow that data or use it for a specific purpose can be challenging. And as discussed by Kitchen, when data increases and more information is received, using an adequate model is an obstacle in itself. Another challenge is the social sciences. Data can be used to measure many of the physical sciences but what about the arts and things that are not as concrete. Can they be measured in the same manner. In terms of individuals and the community, I’m not sure how data science will be used to directly measure individual behavior. Perhaps targeting individuals for specific ads based on recent activity could be an example. However data science could definitely measure the overall community and infer changes/improvements on a larger scale. For example, maybe data science can measure where money is most spent within the community and what locations receive the most activity. This information can allow public officers to better understand the people within the community and their behavior. Or maybe data science could be used to measure where in the community is the most traffic at a specific time or what roads have the most crashes on them. Studying that data could be used for safety purposes and to make changes on road infrastructure. Using scaling and accurate models, data science can more holistically understand how we act as human beings and how we have developed as a society. The applications of that data offer so many opportunities. Companies and the government can use that data to improve the human condition, benefit society, and maybe even predict what changes may happen in the future.