Becoming an Azure Data Scientist Associate: How to pass DP-100 exam
The purpose of this article is to explain what the Microsoft azure DP-100 exam is, the content of the exam, and what I did to pass the exam. I believe this article will guide you with preparing and taking the DP-100 exam.
Prerequisites for passing the exam
- Basic understanding of python.
- Basic understanding of Machine Learning and have worked previously on classification and regression projects.
What is the Microsoft Azure DP-100 exam?
The DP-100 exam (Designing and Implementing a Data Science Solution on Azure) by Microsoft is the exam you need to take to get the Azure Data Scientist Associate certificate, it tests your knowledge of data science, and machine learning to implement and run machine learning workloads on Azure. This is the only exam you need to pass to be an Azure Data Scientist Associate. There is an optional prerequisite Microsoft Azure Fundamentals certification which exam is called AZ-900 exam, I wrote an article on Passing Microsoft Azure AZ-900 Exam with free azure Account and Tutorials.
What is the content of the Microsoft Azure DP-100 exam?
The exam will not cover preview features on the Azure Machine Learning Studio, it only covers features that are generally available. The core content of the exam is divided into the following.
Set up an Azure Machine Learning Workspace (30–35%) — This involves creating an Azure Machine Learning workspace, managing objects in an Azure Machine Learning workspace and managing experiment compute contexts.
Run experiments and train models (25–30%) — This involves creating models by using Azure Machine Learning designer, running training scripts in an Azure Machine Learning workspace, generating metrics from an experiment run, and automating the model training process.
Optimize and manage models (20–25%) — This involves using Automated ML to create optimal models, using Hyperdrive to tune hyperparameters, using model explainers to interpret models and managing models.
Deploy and consume models (20–25%) — This involves creating production compute targets, deploying a model as a service, creating a pipeline for batch inferencing, and publishing a designer pipeline as a web service.
What I did to pass the DP-100 Exam
My background story before I wrote DP-100 exams, I had just passed the Microsoft Azure Fundamental Certification AZ-900 exams in the heat of the Covid-19 pandemic, precisely March 24th, 2020, you can read my article on Passing Microsoft Azure AZ-900 exam with free Azure Account and Tutorials before writing the Microsoft Azure Data Scientist Associate Certification DP-100 on April 7, 2020. I was in the penultimate semester of my masters program in Data Analytics(Machine Learning and Big Data), so I had knowledge of Data Science, please note that the AZ-900 exam is not a prerequisite for the DP-100 exam. There are 4 ways to take the exams:
- At a local test center
- At my home or office
- At a Certiport test center
- I have a private Access code
Following these options, you have a choice of your desired date and time to take the exam. My preference was to take it at home It took me about 3 full days to prepare for the exams, a full day means studying for about 8 -10 hours a day. On the first day I completed the Microsoft Learn Platform tutorial, on the second day I completed the Pluralsight tutorial, and on the third day, I did revision and completed the sample questions on examtopic.com.
To save money as a student, I used only free account and tutorials, but note that you need to understand how Azure Machine Learning designer work, and it requires an enterprise edition of Azure Machine Learning workspace. This enterprise edition requires a paid subscription but the videos on pluralsight are very helpful to learn about the Azure Machine Learning Designer. Please it is also important to understand every line of code in the practical. Details about the free resources I used are written below.
What free resources did I use?
Microsoft Azure Free Account — This is a free azure account that comes with a 12 months validity. $200 free credit is offered for the first 30 days and more than 25 products are accessible for free afterwards.
Microsoft Learn Platform Tutorial on Building AI solution with Azure Machine Learning — There are two training options offered by Microsoft for the preparation towards the DP-100 exams, first is the free online training and the second is the paid instructor-led training. The free online training consists of 7 modules as of the time I took it, it was estimated to take about 5 hours 19 minutes but you might need more hours since you need to go through all the links in the training, understand and run every line of code in the practical section. The code is written in python and connect to Azure using python SDK.
Pluralsight Microsoft Azure Data Scientist (DP-100) Tutorial — It teaches how Azure services work together to enable various parts of the Machine Learning workflow. It consists of 25 lectures estimated to take about 48 hours but since I have an understanding of data science workflow already, I focused more on the Building and Deploying Models in Microsoft Azure section.
Microsoft DP-100 Exam sample Questions on examtopics.com — This website contains about 60 sample questions for free, these sample questions will help you to understand the type of questions to expect in the exam.
What does the DP-100 exam look like?
Price — The exam costs 165 euros (as at the time I wrote mine) since I am in Germany, but the price varies by country. It costs $165 in the USA though $1 does not have the same value as 1 euro. There was a discount for students so it cost me 91 euros before VAT and about 110 euros after adding VAT. There are also discount for Microsoft Imagine Academy program members, Microsoft Partner Network program members or a Microsoft Certified Trainer. You can cancel the appointment 6 business days before the exam day to get a full refund.
Exam language — English, Korean, Japanese and Simplified Chinese
Questions — I had 50 questions although I learn it can vary between 40–60 questions. I had two sections of questions, first is the questions you can review later and the second is the questions you can’t review later. For me, my first 10 questions I could not review them later. To review a question later you need to check the review option. There are different formats of questions like complete the code, repeated answer choices, arrange in the correct order, build a list, multiple-choice single answer, multiple-choice multiple answers, mark review, review screen, active screen, best answer, drag, and drop, etc. Note you are not required to type (input) any answer during the exam.
Duration — the exam lasted for 180 minutes (3 hours) but the scheduled time was 210 minutes(3 hours 30 minutes) that is you have extra 30 minutes to do other things like reading the exam guidelines, taking a survey and a review.
Exam structure and Score — You have the opportunity to select some questions for review, which will give you the opportunity to review and correct them before the final submission. After the final submission, you have another 10 minutes to see all the answers you supplied but this time you cannot make any correction. To pass the exams you need a total of at least 700. Your exam report will be available immediately after you finish. This report contains your total score, pass or fail status and a bar chart showing performance in key areas of the exam. The certificate will be out about 2 days after the exam. You have the opportunity to retake the exam after 24 hours if you fail. You can read more about Microsoft certification exam policy and frequently asked questions here.
I also taught Microsoft Azure Fundamentals for Data Science at the 5-days AI everyday Virtual Bootcamp, this video can be of help. I wrote an article on Passing Microsoft Azure AZ-900 Exam with free Azure Account and Tutorials. I wish you the best as you start the journey in Cloud computing and in your exam. Happy Learning!!!
Originally published at https://trojrobert.github.io on April 13, 2020.