Data Streaming

I'm a senior java and c# developer. I've had a really small exposure to Kafka in my work (for log ingestion), but I wanted to better understand Kafka and streaming application.
It's my first Udacity paid course. Both parts feature several "labs" where you can code in pyhton and use the terminal directly in the browser. That's a nice thing, as you can start applying what you just learn in a matter of seconds.
Overall, I took a month to complete the course (it is advertised for 2 months), by working ~ one hour per day, everyday.

This courses is divided in two : first kafka then spark streaming.

The Kafka part is great. Like, really great. (5/5 without taking the price into account)
The topics (pun intended) are well covered. The labs are well done, even if a little repetitive. The instructor is good and its explanations are clear.
The project is a nice wrap up of all the technologies encountered (kafka, kafka-connect, kafka-sql, faust), with a believable use case. A docker-compose setup is provided so you don't have to install each bit of software and can focus on development.
At the end of this first part, I was feeling like I understood what the Kafka ecosystem was about and how to articulate a solution around it.
The only potential downside was the use of Python : Python is great and accessible, but Kafka is developped in Java and using Python means we can't use Kafka-Stream (we use Faust instead, which is fine).

The Spark part is not so good. (3/5 without taking the price into account)
It's not always the same instructor. The Spark overview is fine (I think thoses lessons are shared with other course about spark). But the specific Spark Streaming lessons are muddier. Sometimes I wasn't sure to understand what was being teached. I get the "idea", but some of the "why" and "how" were lost to me.
In the same manner, the labs are inconsistent : the first ones are done in jupyter with nie "step by step" instructions. But soon, the labs fallback to simple python scrips (fine), but the expected goals (what to modify in the script) were often blurry. I had to resort to the solution files to understand what was expected.
On one occasion, the solution wasn't even working... (due to a simple syntax mistake)
The project was meh. It is nice that it build upon your Kafka knowledge, but the setup was laborious : no docker-compose here, you have to manually install java, kafka, spark, ... The actual coding was shorter. I felt like I was doing a big lab, instead of a project.
And it's a shame : the first project can really be used as a portfolio. Your interviewer just has to "docker-compose up" and he is going. But I doubt anyone would bother running the second project.

The additional side courses about improving your github and linkedin profiles are a nice touch. I helped me find motivation in being more consistent on my side projects.

Bottom line :
- there are some quality issues : code solution not working, painful project setup. The kafka-streaming part need a rework.
- it's pricey. I make a decent living and I have some spare cash, so a few hundreds are ok for me, but I still "rushed" to finish in only a month and avoid paying double.
- it won't make you a kafka or spark streaming guru, but you will feel confortable if you are pushed in a job with thoses technologies. You can build from there.
- when/if the spark part is updated to be the same quality as the kafka part, I will actively recommend the course to some of my friends/colleagues. But not in the current state.

To Udacity staff : please improve the spark part.
Even better, make the spark project a continuation of the first project, that way your students will end up with a "big" streaming application combining kafka, kafka-connect, kafka-sql, faust and spark in their portfolio.

I #highly #recommend this program to all data engineering professionals ( especially those who love streaming of data ).

Whether you're a top-down approach learner or bottoms-up approach learner , both perspectives are taken care with pragmatic methodology in this professional course.

You get to do two projects as part of completing this course , which i feel is the highlight (and the best ) part of this course.

Trust me , you'll atleast go for project submission 2-3 times.

Watch out for review comments from professionals, these comments will not only help you in completing the projects (short term) but also help you in professional your career( long term).

This is my first udacity nano degree coure , and i just enjoyed it to the core. Thanks to #udacity for creating such a wonderful course.