The terms Data Science and Machine Learning are often confused, even used interchangeably by some. In this article, however, we dissect the whole topics to help you in understanding the difference between Data Science and machine learning.
History of Data Science
To put everything into perspective, we’ll dig just a little bit into the history of Data Science.
As early as 1962, the field of statistics was experiencing a dynamic shift. One John Tukey said: “… as I have watched mathematical statistics evolve, I have had cause to wonder and to doubt…I have come to feel that my central interest is in data analysis…” At the time, statistical analysis could be presented in hours as opposed to days or weeks. He refers to when computers had just come.
Fast forward to 1999, there was data everywhere. Most businesses were looking for ways to make better, informed decisions on consumer patterns, preferences and behavior.
Data science got its true place. John Zahavi, at the height of data explosion said:
“Scalability is a huge issue in data mining… Conventional statistical methods work well with small data sets. Today’s databases, however, can involve millions of rows and scores of columns of data… Another technical challenge is developing models that can do a better job analyzing data, detecting non-linear relationships and interaction between elements… Special data mining tools may have to be developed to address web-site decisions.”
This snippet of history is key in demystifying the hot topic of Data Science vs Machine Learning as we are about to see.
Now what is Data Science?
Data Science is an interdisciplinary subject that employs scientific theories, methodology, algorithms and machine learning, to mine cluster, analyze and produce useful insights based on the primary complex data.
Data science is a broad topic that covers the entire data mining, statistics, algorithms and programming. In fact, most data scientists are not just statisticians but are also well versed in popular programming languages and coding.
It involves the whole idea of data generation, processing and even storage. Talking about data storage; this is a field that has undergone tremendous transformation. In 2011, Pentaho CTO, John Dixon coined the concept “Data Lakes” which enables data scientists to store large, non-linear data as opposed to “Data Warehouses.” Data Lakes store all sorts of data regardless of their nature, classified, linear, non-linear, SQL or NoSQL.
Data Science covers the topics as:
- Automated data-driven decisions
- Machine learning and automation
- Data Visualization
- Data Engineering
- Data Integration
- Distributed architecture
- Deployment in production, and more.
What then is Machine Learning?
As we’ve seen with Data Science, Machine learning is a subset of Data Science. Tom Mitchell’s simple definition of Machine Learning is: “A computer program is said to learn from experience E with respect to some class of tasks T and performance measure P if its performance at tasks in T, as measured by P, improves with experience E.”
Machine Learning derives its origin from Artificial Intelligence with the main intention being to automate as many processes as possible.
Machine Learning, therefore, learns from data, identifies patterns and makes informed decisions based on experience, with minimal or no human intervention. These are the supervised learning and the unsupervised learning.
Data Science and Machine Learning
Like Data Science, Machine Learning employs mathematical models and algorithms in computers to help them make autonomous decisions. The term “learning” in Machine Learning means you use some set of data to feed into the mathematical models and computer algorithms. It is this data that makes the machine make informed decisions.
Conclusion
The main difference between Data Science and Machine Learning is this: While Data Science is chiefly concerned with drawing inference and solving statistical and analytical questions, Machine Learning is concerned with optimizing performance.
In understanding the difference between Data Science and Machine Learning, we must have in mind that they heavily borrow from one another even as one is the subset of the other.
