Akansha Yadav maps out ten skills for becoming a data scientist in 2024: math, Python, SQL, exploratory data analysis, machine learning, deep learning, MLOps, cloud platforms, big-data tools, and portfolio building. She recommends Kaggle competitions and research papers for real-world practice.
Ladies and gentlemen, hello everyone, welcome back to my YouTube channel. In today's video, I am going to take you to a roadmap using which you can be a data scientist, and I'm going to tell you each and every skills that are required for this role. So let's get back on the screen and see what are those. So I have divided this roadmap into the 10 different parts. So number 1 is the maths, then you have to learn about statistics, along with that you have to learn about coding, then EDA, and then machine learning, following up with the deep learning concepts, and then MLOps, case studies, and some basic tools you have to learn, and then at the end you have to build a complete portfolio. So in mathematics, you have to learn about the linear algebra, then you have to follow up with the calculus, probability, and statistics. There are multiple things in the statistics that you have to learn, which is— which includes the confidence intervals, regression analysis, and ANOVA analysis as well. In programming and in the coding part, you have to learn about the Python programming language. You have to master in Python programming language, which includes Data structures as well. You have to learn some libraries as well, like NumPy, Pandas, Matplotlib, Seaborn. And there are other libraries also, like sklearn and TensorFlow, which will— which you will be learning in the machine learning and the deep learning concepts. So then you have to learn about the SQL, which is query-based database. Like, this is the database management tool. You have to learn about this. Then you have to follow up with the exploratory data analysis. Exploratory data analysis, you have to learn about data cleaning, data visualization, descriptive statistics, feature engineering. Once you are done with all these skills, you have to learn some basics of machine learning. Not only the basics, but yes, you have to create some models as well, because you are going to be a data scientist where you will be required to make decisions using the machine learning models. So you have to learn about the supervised learning, unsupervised learning, model evaluation, hyperparameter tuning— like how to, uh, make your model capable of giving you the best accuracy. Following up with the deep learning, in deep learning also there are multiple concepts. Deep learning is basically based on the big data concept, so you have to learn about some deep learning models as well, like neural network, deep learning libraries also you have to cover, like TensorFlow and PyTorch from Pandas, which will help you build and train deep learning neural network models. Then you have to learn about the conventional neural network, which is CNN, uh, CNN, and then you have to learn about RNN, and then traffic learning, transfer learning as well. Then you have to learn about the MLOps. MLOps is nothing but the machine learning, uh, operation. And in this you have to learn about how to deploy your model. There are multiple frameworks in Python which you can use to deploy your model. One of them is the Flask, and then, then we have the FastAPI as well. You have to monitor your model to understand the techniques from— you have to understand the techniques for monitoring your model. Then you have to learn about the scalability, and along with that you have to learn some— not learn, just go through this case study so that you will understand that what industry's demand is and what they are working on and why do they need data scientists and why they are going to hire you as a data scientist. So you have to read some case studies, you have to work on some real-world projects, uh, which is in the different domains. You can choose your domain and as for that you have to work on it. So I would suggest like if you are in the finance, go with the projects or go with the case studies of finance. You can join the Kaggle competitions as well. You can read some research papers. Then after that, you have to learn some data science tools, like cloud platforms. There are multiple cloud platforms which will help you to deploy your models, like AWS, Google Cloud, or Azure, for data storage, processing, and analysis. You can learn some big data technologies as well, like Hadoop, Spark, or Hive. For data engineering also, you have to learn some process, like how to do data pipelines, how ETL processes works, and how— what is data warehouse. Once you are done with all these things, you have to build your portfolio. To create a portfolio, you can use a GitHub as well, or you can create a personal website which do not require any domain from GoDaddy. You can just use a Netlify or something like that to deploy your personal website. You have to keep engaging with the data science community to excel in this field. So if you want this PDF onto your WhatsApp, you have to join my WhatsApp channel. I have already created one WhatsApp channel which like I have added the link in the description. You can go check that out. I have uploaded n number of resources for data science role, for machine learning role, or for data analyst roles. So for this PDF also, you have to follow me on WhatsApp. So I hope you like this video. Don't forget to comment your views on this. I have made it really quick to just make it shorter. So bye-bye.
