One of the most critical factors for any data scientist (using the term in the most broad sense) is to keep learning over time in order to excel the skills that we know and to discover new knowledge and techniques. I haven’t done any online course in quite a while, and it is about time I do it. For this reason, and due to my lack of free-time, I have done a bit of research to select the best possible online (and free) course that is starting in the near future. However, over the journey of finding such course, I have come across several courses that look enlightening and exciting that I will definitely do if I had more free time.
Going against common sense, I will start from the end. The course I have chosen is Mining Massive Datasets from Stanford University (hosted in Coursera). The main reasons behind this choice are the syllabus (with a huge overlap with the technologies and methods we are using in Signal), a great university and an even better set of lectures. Also, the timing (starting at the end of September) and duration (10 weeks) suits my schedule. The main negative factor is that the estimated effort per week (8-10 hours) is slightly more than I would like to invest in the course at the moment, but I think it is worth the extra effort. The rest of courses that I have considered are:
- Learning From Data. 25th Sep 2014 (10 weeks). 10-20 hours/week effort. edX.
This seems like a very good course to start in the Machine Learning field. However, I am looking for something more advanced. I would probably have a look at the last lectures, especially the ones about SVMs.
- Networked Life. 1st September (7 weeks) 1-3 hr/week. Coursera.
Sometimes we narrow our perspective by choosing specific courses (i.e., related to Machine Learning) and we ignore other related disciplines or fields. This course appears to give a general overview of graphs and network analysis from completely different application perspectives such as social media, finance or even biology. In addition, it claims that it requires only 1-3 hours a week. If I had more time, I will also join this course. In any case, I am certain that I will watch some of the lectures.
- Data Analysis and Statistical Inference. 1st September (10 weeks) 8-10hr/week. Coursera.
Probably the most mathematically-focused course of all the ones in this list. Statistics is a very important part (one that we tend to forget about) of Machine Learning and Information Retrieval. This course covers some of the foundations in data analysis and statistics.
- Mining Massive Datasets. 29th September (7 weeks) 8-10hr/week. Coursera.
This course explains several concepts related to mining massive amounts of data. The syllabus shows quite diverse tasks from MapReduce to de-duplication, classification or recommender systems. Furthermore, the concepts explained in this tutorial are very pragmatical and applicable in several scenarios.
- Data Science Specialisation. 1st September (42 weeks) 3-6hr/week. Coursera.
This combination of nice courses from John Hopkins is the first coursera specialisation that focused on data science. It focuses on applied Data Science and this is clearly illustrated in the syllabus. My humble opinion is that this is a lot of time (almost a year) to spend learning this concepts. However, this is only true for people who already have knowledge and experience in Data Science, Machine Learning or any related discipline. This course is probably very recommendable for people starting from scratch (probably from a software developing position) who want to apply some machine learning concepts.
Thanks to this little review of the course offerings I have rediscovered class-central, a centralised repository that allows to easily search the courses from the main e-learning platforms (e.g., edx or coursera), as well as independent entities. It is definitely worth having in your bookmarks. Also, this code will allow you to download lectures from coursera automatically. This is extremely useful for people like me who spend a lot of time in the tube and the train without connection.
As mentioned in the beginning of the article, this list only list free courses that are about to start. There are many amazing courses that are either non-free or that have started (or even finished) already such as the worldwide famous Machine Learning by Andrew Ng. In addition, I am sure I have overlooked and not found other courses that fit my description. Please drop a comment if you have any other suggestions.