Earlier every Computer Science student wanted to pursue a career in the data science field. The field also attracted students from many other educational backgrounds. While the hype around data science still exists, the job profile isn’t readily available for all.
The past two decades have been a revolution for the data science community. The developments in the past two decades are phenomenal and have taken everyone by storm. The applications of data science have amassed all the industries and had readily increased the demand of data scientists.
But the trends are changing nowadays. The demand is no longer the same as before. Even if there is a demand for data scientists, either people lack the skill set or the experience. I have tried to list all the potential reasons I think of when I see the community losing its charm.
Once out of universities, people want to kick start their careers as a data scientist but the job offerings require a minimum of 2–3 years in most of the cases. People are not able to directly start as a data scientist and have to start their career in different profiles.
Companies are not ready to invest their time on new incoming talents instead they want the ones with the excellent skill set and experience in the field. While almost all the tech companies have their own Data Science departments, the others which do not have one, need a person with a lot of experience in this field to start one.
There’s only one way I think that could help them is doing internships while they study and gain experience to meet the demands of the companies.
Another major reason is that data science enthusiasts nowadays do not know the difference between different job profiles available in the field. All of them want the ‘Data Scientist’ title without knowing what the actual work of a data scientist is. They mistakenly consider Data Analyst, Business Analyst and Data Scientist as being similar profiles.
Without knowing what they want to work upon they apply into roles they aren’t a perfect fit for and end up empty-handed.
People directly start working on learning algorithms, ways of tweaking the data but what they do not consider is the Math behind the algorithms. With average programming knowledge and knowledge of the Machine Learning algorithms, they think they are ready to face real-world problems.
People usually ignore statistics, the actual hard work just because they do not find it interesting. Data science is one such field where the development isn’t stagnant. Natural Language Processing has seen some massive developments in the past 2–3 years. One has to keep themselves updated with the state-of-the-art models.
People also find data science easy because they haven’t worked on real-life data. All the years that they have spent learning, they have worked on structured data or some pre-processed data that was made available for people to learn.
On the other, almost 99% of the data in the real world is unstructured. Data scientists need to spend most of their time pre-processing the data so that they can extract something meaningful from the data.
As soon as tech giants Google, Microsoft launched AutoML, it shook the aspiring data scientists. Companies’ interest and their curiosity grew into AutoML, while data scientists fear losing jobs.
AutoML is automating the process of applying machine learning to the datasets. AutoML can preprocess and transform the data. It can cover the complete pipeline from working on raw data to deploying machine learning algorithms.
AutoMLs are good at building model but when it comes to preprocessing, they cannot outshine humans. The major work of the data scientist lies in pre-processing the data. It is clear that as of now AutoMLs cannot replace human data scientist.
Although the fact that AutoMLs reduce the costs cannot be overlooked. The average annual salary of data scientists in the US is around $120k whereas the cost annual cost incurred by Google and Microsoft AutoMLs is somewhere around $4k to $40k.
Though the effectiveness of data scientists at pre-processing data cannot be denied because the data in the real world is highly unstructured and requires a lot of pre-processing.
There so much to learn and no one is willing to do the hard work. It is difficult for someone to start with the basics and excel in this field. This would take a lot of time and people need to be patient. There is a lot of scope in this field but the lack of people with the actual skills needed is snatching the title of most promising job away from Data Science and people are walking away from it.
