Data Science: Trusted Relationships & Scientific Exploration
Posted by Tech in Motion
Tech in Motion Orange County will explore Data Science through a tech panel on Wednesday, July 20th at 6 p.m. at The Cove at UCI Applied Innovation. The event will include speakers across different tech sectors applying Data Science today. Below is a guest blog from one speaker on this panel. Learn more and register for this free event today! REGISTER HERE.
Written by Dave Herman Vice President of Applied Science at Payoff, a leading financial wellness company, using science, psychology and technology to help their members reinvent their relationship with money and accelerate their journey to financial well-being. You can learn more about Payoff's Science team here.
The new field of data science is rapidly enabling us to link massive amounts of seemingly unrelated ideas, and pool that new knowledge to potentially predict outcomes of virtually any situation.
How do data scientists do this?
To a great extent, we do what scientists have always done: We look for repeatable answers to the questions that surround us, by finding meaningful patterns in data. In the context of psychology, we look for patterns that allow us to understand and predict human behavior, behaviors that are expressions of beliefs, habits and desires.
The process of finding relevant patterns in these big data sets has been exponentially advanced by the adaptation of more powerful computational infrastructures and machine learning algorithms.
As data scientists, we use our knowledge of algorithms, computational tools and a wide range of subject matter expertise from natural language processing — from the digital interpretation of human language to behavioral psychology to neuroeconomics. Through a combination of scientific method and machine learning, we develop an evolving and increasingly more powerful set of insights and methodologies to explore large data sets in the big-data clouds.
Fundamentally, data science is about finding patterns and information, hidden within data. Imagine a tiled mosaic, in which each tiny tile represents a unit of data. At close range, we’re limited in the information we can gather, focused around the attributes of individual tiles. But only when you absorb and compute on the entirety of the data, can you see the deeply insightful patterns. When we stand back, take it all in and see the collection of images and story at full scale, that’s when we learn from big data.
Someone who understands the sequential process of data science explores the data structure that holds this mosaic together using algorithms and their knowledge of math, science and statistics to guide this exploration, finding patterns. Without patterns, the data is useless. Once a pattern is found in the data, it becomes very meaningful and we’re able to learn important insights potentially applicable to wide segments of people.
Our digital behaviors, including “likes” on social media, clicks to buy products, and measurements of physical and mental activity (Fitbits, accelerometers, questionnaires, MRI and EEG) all offer views into the core of human individuality, and allow data scientists to see humanity on a very large scale.
Data Science Uses the Most Powerful Machine Around
Working with information to determine how to tell a story at this vast scale, data science uses the most powerful machine around - the human brain - to develop algorithms that allow computers to take over the storytelling and point the information in the right direction.
Data can’t tell a story without a digital brain to organize the data and a human brain to contextualize it into human experience. Much of our digital lives is powered, at some level, by the decisions of people about what patterns matter, and what algorithms and experiences are needed and intended. It’s where digital interactions get the contextual IQ that transforms data into information, followed by experiences that actually connect with humans.
Think of an algorithm as a means to make the complicated digestible, as it creates the filters/structure and statistical summarizations that enable specific points to come into focus. And like with everything else in life, what you get out of it is a result of what you put into it. While an algorithm creates the matching, its value is a result of it taking into consideration things we really care about.
So, what do we really care about? As it turns out, almost everything where human behavior is concerned. This is science, after all, and the iterative, empirical process of study can be applied to everything we humans do.
As we increase the application of applied data science across areas of focus, what matters is who or what is behind the algorithm. Data is one thing, but telling a story that captures the value in the data in a way that earns people’s trust is something different altogether.
Take for example, the “stable marriage problem.”
Say you have a large set of people X, who each desire someone from set Y, and vice versa. Here’s the task ahead of you, the ultimate cupid: Is it possible to optimally assign pairs where they are all “stable”? (Mathematically speaking, “stable” here means there are no people in both sets who would rather be with someone other than their current match.)
As it turns out, there is a smart algorithm for solving this problem. However! Creating the algorithm is just half of what a data scientist must do. They also need to know what information matters and what makes for a good match. The algorithm lets us do the matching, but it’s the features, traits, assessments and real-world knowledge that make that match effective.
Is that a little too George Orwell, dystopian nightmare for you, you say?
Let’s look at eHarmony, for example, which is powered by an algorithm developed in part by Dr. Galen Buckwalter, co-author of this article. We all know couples who met online — people who would likely never have found each other otherwise. The most important decision in an adult’s life — who to choose as a partner — has been improved, thanks to applied data science and the intelligence of the human minds who pointed their emotional intelligence in the right direction, toward love.
Could there be a more profound use behind brains and algorithms working together?
The data — a long list of very specific attributes selected as appealing by an individual — leads to a match with someone who has chosen many of the same attributes. Then the humans make a date and get to decide if the connection will become more than an awkward dinner, albeit with a stranger with whom you share many attributes.
Some may call it unromantic and perhaps lacking in the serendipity we tend to assign to romance, but the scientific process is exceptionally effective, as is evidenced by the thousands of people who engage in this experience every day.
Data science in action is an algorithm that guides a user toward choices they would probably never have found without it. Given the seemingly infinite possibilities of a connected world with more than 7 billion people in it, matching us with what we want and need becomes increasingly valuable every day, as more and more data is collected.
This post was adapted from the original on Medium.com, "Data Science, Trusted Relationships and Scientific Exploration," which you can read in full here.
Read more about data and technology:
- When Everything Becomes Data
- Revolution: the Cloud & Multi-Channel Communication
- 8 Tech Companies That Will Revolutionize Your Life