Microsoft puts a lot of efforts into technologies that are related to data, no matter whether you’d call it “Data Analytics”, “Data Science” or use the classic term “Business Intelligence (BI)”. For a company of this size it is no surprise that they have several offerings regarding education. When it comes to certifications, there are also two main “lines” in the meantime: the classic Microsoft Certification Program (MCP) with Microsoft Certified Solution Expert (MCSE) / Microsoft Certified Solution Associate (MCSA), and the rather new Microsoft Professional Program.
I’m looking at “Data” only, so I narrow this post down to the relevant data certifications. These are:
- Microsoft Certified Solutions Expert for Data Management and Analytics
- Microsoft Professional Program for Data Science
While the MCP is rather focussed on technical skills using Microsoft software like SQL Server (to name the main software piece for “Data” in general, at least on premises), the Professional Program adds more conceptual skills and brings them together with the technical skills. These conceptual skills go from mathematical and statistical basics over ethics/law to data science methods in the narrower sense.
Microsoft Professional Program for Data Science
Microsoft manages to bring these topics together into one platform. They are using edX as a provider for the courses. edX is an online learning destination that was founded by Harvard University and MIT, so they host lots of courses, not only the Microsoft ones.
The current Data Science Program consists of the following courses:
- Course 1: Introduction to Data Science
- Course 2a: Analyzing and Visualizing Data with Power BI
- Course 2b: Analyzing and Visualizing Data with Excel
- Course 3: Analytics Storytelling for Impact
- Course 4: Ethics and Law in Data and Analytics
- Course 5: Querying Data with Transact-SQL
- Course 6a: Introduction to R for Data Science
- Course 6b: Introduction to Python for Data Science
- Course 7a: Essential Math for Machine Learning: R Edition
- Course 7b: Essential Math for Machine Learning: Python Edition
- Course 7c: Essential Statistics for Data Analysis using Excel
- Course 8a: Data Science Research Methods: R Edition
- Course 8b: Data Science Research Methods: Python Edition
- Course 9a: Principles of Machine Learning: R Edition
- Course 9b: Principles of Machine Learning: Python Edition
- Course 10a: Developing Big Data Solutions with Azure Machine Learning
- Course 10b: Analyzing Big Data with Microsoft R
- Course 10c: Implementing Predictive Analytics with Spark in Azure HDInsight
If there are several letters a, b, c for a number, only one of the courses has to be taken. That means that there are 10 courses that have to be taken successfully, plus the so called Capstone Project at the end of the course. The capstone is a challenge where you have to solve a real world problem and develop a machine learning solution. At the time I’m writing this post I have successfully taken all the 10 courses, but not yet the capstone. I’m looking forward to it (starting in October of this year). Here is a very good statement from someone who has finished the capstone: Koen Verbeeck: MPP – The Finish Line.
The courses can be run at ones own speed, but they are limited to three months. The time that is needed to go through one course can differ very much, depending on skills you already have. Additionally I found some courses to be easier and some more difficult – even if I had similar previous knowledge. I would compare a course with an intensive one week of training. Doing it in the evening plus on weekends makes it realistic to take one course in one month or a shorter period of time.
All the courses can be taken for free. However, to earn the Microsoft Certificate at the end, every course must be upgraded to a “validated” course which costs 99 USD per course. Thus the whole Microsoft Certificate is about 1000 USD. It is possible to start with one course and then decide later to go for the upgrade (even once you have finished the course), however because of the time limit you should only start the courses that you intend to finish in the current quarter.
The courses consist of explaining videos, example data, additional material like PDFs, exercises and so on. Additionally there are knowledge checks (often multiple choice) of different difficulty levels. Some are easy, some are not so easy, but you have several attempts, sometimes you have only one attempt, and one course had a time limit. On the whole, everything was manageable for me quite well. The statistical basics where the area which was most difficult for me, even if it was just a refresh of what I learned in University. Probably it was just the least fun topic.
Microsoft Certified Solutions Expert for Data Management and Analytics
This approach is probably much more known than the Professional Program. And I guess it is widely accepted in the industry.
Of course the certifications have changed over time, but the basic concept has stayed more or less the same. A certification consists of several exams. It is possible to just take the relevant exam, but there are also courses that prepare for the exams. Many training companies offer these courses. They have unique numbers, so one can know which one to take to be prepared for the exam.
Of course the certification for Data Management and Analytics is very much SQL Server focused. The version is relevant. There are old certifications on 2012 level (I ignore them here), and the current ones are on SQL Server 2016.
The complete certification is the Solutions Expert (MCSE). The “smaller” certification is the Solutions Associate (MCSA). The following table shows all the exams/courses that belong to the area of Data Management and Analytics (as I wrote: 2012 ignored). The MCSA always consists of two exams, so two rows in the table.
For some exams there is official Microsoft literature. There are links to the books given in the table. Actually they link to Amazon as “affiliate links”, that means that I earn a provision if you buy one something at Amazon after clicking on one of the links.
To obtain the MCSE certification, you must first get one MCSA certification, then one additional exam belonging to another MCSA. So you don’t even have to take 2 MCSA certifications, nor do you have to touch more than 2 different MCSA areas. There can be quite a big spread between two MCSEs, one obtained by doing the classical SQL Server route, one doing the Azure/Machine Learning route.