Data science is a branch of science that combines scientific knowledge and people’s knowledge about the question of interest to get insights from data. Data science has an important role in businesses because it helps them determine whether or not doing something is worth doing. The scientific knowledge is knowledge from mathematics, statistics, and computer science.
A data scientist is a professional who gets data, programs code, and applies mathematics, statistics, and computer science to create insights from data.
This is how science is typically working. A scientist observe the world, then make a guess (i.e.,a theory) about how the world might be working. Then a scientist does all sorts of experiments, and try to prove, disapprove or modify the guess.
The data science works in a similar way. A data scientist observe the world from the data, then make a guess (i.e.,a theory) about how the world might be working. Then a data scientist does all sorts of data mining, and try to prove, disapprove or modify the guess.
The data science is try to understand how the world is working as much as possible. However, we rarely know how the world is exactly working. This is similar to the “Blind men and an elephant”. We are tying to understand how the world is working (the elephant). We get some data generated from the world (legs, tails, nose, etc of the elephant). Then, we make a guess on how the world is working (how an elephant looks like).
There is always limitations in data. The insights we draw from the data is no better than the data (Garbage in, garbage out). The data we get might be too old (outdated), or the data does not represent the population (biased data), or the data does not have the level of details (not granular enough), or the data have too much details (dimension problem), or something will change in the future (the past does not represent the future).
While data science has an important role in businesses, another branch of related field that helps businesses is machine learning. Machine learning is to find a pattern in data (build a model based on data), in order to make predictions (decisions). There are two situations in machine learning.
One, the data has both related information (inputs) and the information of interest (outputs). The goal is to identify a pattern that links inputs to outputs (Supervised learning). For example, take a look at customer profile (input: e.g, credit score, income), and identify who are more likely to go default (output: e.g., default or not).
Two, there is no separation between inputs or outputs in data and the goal is to identify patterns in data (Unsupervised learning). For example, take a look at customer profile (e.g, age, gender, bought some book, go to some restaurant), and identify if we can group customers together by a pattern.
Now, we know what is data science/machine learning, next, let us move on how it looks like at work.
Amazon job on data scientist/machine learning: creating recommendation solutions to understand customer shopping patterns
Facebook job on data scientist/machine learning: which ads to display to users
Chase job on data scientist/machine learning: Design and develop machine learning models for the card business throughout the customer lifecycle (e.g., acquisition, account management, transaction authorization, collection)
In conclusion data science and machine learning are both really important to businesses becoming more successful.
If you like my post, click Like button below, leave a response below, or share it to friends. For more details on data science and machine learning, please refer to my other posts as well as online community of data scientists and machine learning practitioners:
Kaggle, a subsidiary of Google
GitHub, a subsidiary of Microsoft
Towards Data Science, a Medium publication sharing concepts, ideas and codes
Leave a Reply