The Best Online Courses for Data Scientists
The Data Scientist profile is one of the most demanded profiles in the labor market. At the same time, the Data Science toolset is becoming more diverse and the skills demanded are broader. Luckily, for those trying to take the first steps into Data Science or mastering techniques, there are many excellent online courses.
After publishing a list of recommended Masters in Data Science in Spain this summer, we have decided to fill up the gap for those who prefer the flexibility of online, and domain-specific courses. This list has been curated by some of our data scientists and it is based on their experience. It covers Machine Learning fundamentals, as well as specialization in Deep Learning, Natural Language Processing (NLP) or probabilistic modeling.
We have grouped courses depending on specialization required. The list is obviously biased, but we can assure you the Data Scientist at BBVA Data & Analytics that have collaborated accumulate an extensive knowledge of the best training the internet can offer. Some of these courses serve as an introduction to certain disciplines and some others are part of larger training programs that can be completed with more courses.
In this first level, we include online courses that provide a basic ground to get started in data science. They offer a first approach to the languages and tools used, as well as some more complex disciplines such as Deep Learning.
- Data Science Specialization. Coursera. Course created by Johns Hopkins University. This comprehensive introduction to Data Science is formed by a set of 10 courses of 4 weeks each. This course is a great way to start in Data Science and to obtain a general understanding of the basics of Data Science.
- Deep Learning Course. Lazy Programmer. This course covers every important theory and practices for those getting into Deep Learning. In the process of learning this increasingly important field of Machine Learning, the student will reinforce important Data Science principles. The plain language in which the concepts are explained is one of its strong points.
- Machine Learning Foundations: a case study approach. Coursera. With 30 to 48 hours, this program by the University of Washington will introduce use cases of basic Machine Learning problems, such as logistic regressions, forecasting, and classification. It is the first of a four-parts Machine Learning Specialization program.
Basic concepts of Machine Learning and Deep Learning continue to be addressed, but the specialization is higher in this second level.
- Machine Learning. Coursera, 11 weeks. This course is taught by AI luminary Andrew Ng and Stanford University. In the 11 weeks of the program, you will go from a basic level of Machine Learning understanding to a solid level both in practice an theory approaches. The course is ideal for those looking to reinforce that maths behind Machine Learning. The dedication and level of mathematical abstraction require a bit more dedication than a normal course.
- Text Retrieval and Search Engines and Text Mining and Analytics. Coursera, 20 hours each one. Taught by the University of Illinois. This course is specialized in Data Mining and is a good introduction to classic techniques of Natural Language Processing (NLP), text analytics, search engines, and content-based recommender systems. The program will offer you an introduction to a vectorial model of documents, TF-IDF techniques, evaluation of classifiers and search engines, sentiment analysis, document clustering, topic modeling, and visualization.
- Practical Deep Learning For Coders. 7 weeks. Put together by the USF Data Institute and taught by Jeremy Howard (Kaggle’s Grandmaster #1). The syllabus can be covered in seven weeks and is divided into two parts. The focus is practical and covers basics such as embeddings, structured Deep Learning, collaborative filtering, NLP, and event Generative Adversarial Networks, considered by Facebook’s Yann LeCun the most important recent breakthrough towards AI.
The courses shown here are aimed at professionals who want to expand their training so that more specific topics and more technical complexity are addressed.
- Probabilistic Graphical Models. Coursera, 2 months. Created by Stanford University. The course is divided into three parts (Representation, Inference, and Learning). The program is directed to those looking for specializing in graphical probabilistic models (Bayesian Networks y Markov Networks) with applications in Machine Learning and decision modes. The strong point of this course is that it asks you to get into detail (applying Matlab and Octave), so the theory is well understood.
- Deep Learning by Google. Udacity, 3 months. This advanced course covers in three months all the necessary steps to become a Deep Learning professional.