Will Data Science be extinct in the next 10 years?

There is much talk that Data Science will be dead in the next 10 years, the main reason, the #AutoML tools.

Data science has been around for a long time, we ourselves as humans constantly observe the world (collect data) and create models that explain our observations.

The only difference today is the volume of data that has grown exponentially with the development of the digital age, not to mention the future that will require creative ideas to combine these new volumes of data.

The AutoML will not solve the problems, they are only tools that will help speed up the testing and implementation of the different algorithms. The ability to get the clean data in a model is not trivial at all.

AutoML creates the delusion that Data Science is only Models, #AndrewNG recently mentioned about the importance of focusing more on data and not models.

Building the model is easy, the difficult part is in:

- Know what tool to use or what methods do not work well
- Know what steps will improve performance
- Know what compensations are important in a given problem.
- Have the knowledge and ability to link the above with the general objective.
- Have the communication skills to interact with experts in the domain.

All of these skills require real-world work, take time, and the learning journey is cognitively demanding.

Let me show you how ridiculously trivial it is to implement the most popular machine learning algorithms these days.

Using Scikit-learn, we can implement decision trees with the following two lines of code:

from sklearn import tree
tree.DecisionTreeClassifier.fit (X, Y)

We can implement Support Vector Machines (SVM) with the following two lines of code:

from sklearn import svm
svm.SVC.fit (X, y)

Can you see the pattern? All we have to do is rename the function and there you have the model. Real data scientists aren’t sitting around and re-implementing these algorithms from scratch. They will end up using a mature library like Scikit-learn or Tensorflow in the industry.

Data science is science for a reason, it is about solving problems that require creative and ingenious solutions. The use cases of data science will only increase over time, this because more and more data will be collected, and with the help of greater computational capacity, much more complex operations will be implemented.

What to do if you are new and you want to enter this world of data? The first thing is, DO NOT worry about whether Data Science will be extinct soon, it will only distract you from enjoying the learning journey. If you continue your work tackling projects from harvest to model implementation, you will be on the right side beyond 10 years.

#datascience #ml #autoML #extinct #model

Source: https://www.kdnuggets.com/2021/06/data-science-not-becoming-extinct-10-years.html