Coursera provides one of the longest-established online data science educations, through John Hopkins University. It isn’t completely free – if you can afford it, you are expected to pay a course and certification fee – but this is waived for students who don’t have the financial resources available.

Comprised of 10 courses, the specialization covers statistical programming in R, cluster analysis, natural language processing and practical applications of machine learning. To complete the program, students create a data product which can be used to solve a real-world problem.

Also from Coursera, this course is provided by PwC so unsurprisingly focuses more on business applications than theory. It covers the spectrum of tools and techniques which are being adopted by businesses today to tackle data challenges, and the different roles that data specialists can fill in modern organizations. Students are also tutored on selecting the best tools and frameworks for solving problems with data. The four-week course concludes with a task involving deploying a data solution in a simulated business environment,

This course is provided by Microsoft and forms part of their Professional Program Certificate in Data Science, although it can also be taken as a stand-alone course through EdX. Students are expected to have an “introductory” knowledge of R or Python – the two most popular languages for data science programming at the moment. Subjects covered include probability and statistics, data exploration, visualization, and an introduction to machine learning, using the Microsoft Azure framework. Although all of the course material is free, students can pay ($90 in this case) for an official certificate on completion.

This course is provided by Microsoft and forms part of their Professional Program Certificate in Data Science, although it can also be taken as a stand-alone course through EdX. Students are expected to have an “introductory” knowledge of R or Python – the two most popular languages for data science programming at the moment. Subjects covered include probability and statistics, data exploration, visualization, and an introduction to machine learning, using the Microsoft Azure framework. Although all of the course material is free, students can pay ($90 in this case) for an official certificate on completion.

IBM provides a number of free online courses through its portal formerly known as Big Data University and now rebranded as Cognitive Class. This program covers data science 101, methodology, hands-on applications, programming in R and open source tools. Collectively they should take around 20 hours to complete although those with prior experience of computer science will probably progress more quickly, whereas complete beginners may take a little bit longer.

This course focuses on machine learning and is delivered as a series of video lectures along with homework assignments and a final exam. As well as an overview of how computers “learn”, it goes into depth with the mathematics (students are expected to have a working knowledge of matrices and calculus, so this one isn’t for complete maths newbies).

Dataquest is an independent online training provider rather than being affiliated with a university like most of the others here. It offers free access to much of its course materials although you can also pay for premium services which include tutored projects. It offers three paths – data analyst, data scientist and data engineer, and with endorsements from Uber, Amazon and Spotify it looks like a good way to get a feel for whether or not you will enjoy studying data science, without spending money.

KDNuggets is a well-known business and data science website and it has compiled its own free data mining syllabus. There are modules on machine learning, statistical concepts such as decision trees, regression, clustering and classification (see my data science glossary for an introduction to these terms) as well as an introduction to practical implementations of the technology.

Rather than being offered by an organization or institution, this course is comprised of a collection of open-source materials and resources, available freely online. Subjects covered include natural language processing of the Twitter API using Python, Hadoop MapReduce, SQL and noSQL databases and data visualization. It also includes a grounding in the algebra and statistics needed to understand the fundamentals of data science. Of course there is no certification but the program can be completed at your own speed and works great as a gateway to the wealth of information on data science available online.