Subscribe to DSC Newsletter

Relevance & need of Python programming in Data Science and Analytics, Machine Learning

I understand that Python is the basic stepping stone in Data Analytics programming side & I must learn upto certain extent of Python (if not expert) to make a career in Data Analytics. However, recently I observed, in Azure Machine Learning Studio, entire machine learning coding job has been simplified just as  module drag & drop. With good understanding on Data Science-machine learning,one can easily manage the modelling with zero knowledge on Python (or any other language) coding. That is what I understood however being newbie in this area I am not confident if I oversimplified things & my understanding is correct or not. Is Python coding knowledge still required? Am I missing something? Even in any Data Analytics course, I see a special placeholder is kept for Visualization using Python (like graph plotting), but there also I see Tableau, MicroStrategy PowerBI or Qilkview which are much easier to use & I don't know if I will still need to use the visualization library & code from Python to visualize my model output or data pattern. Could any of you help with your expert opinion ? 

Tags: Artificial Intelligence, Big Data, Data Analytics, Data Science, Machine Learning

Views: 386

Reply to This

Replies to This Discussion

Great question, I've thought about this myself!

As someone who is NOT a data scientist (full-time digital-marketer, part-time data science student), so speaking solely from a practical, logical perspective:

If a marketer asked me for a job and told me, "I can get the same results as anybody else, I can get the most efficient CPA, drive leads, write copy, etc. ...but ONLY with XYZ tool." I would almost immediately (probably unjustly, but I'm human) write them off as a grunt worker with very little value as a marketer.

Working with different tools, platforms, formats, etc. provides valuable experience, increased insights, and new perspectives. It's the difference between being a chauffeur of knowledge and being a master of your craft. If you resign yourself to a plug and play tool, your effectiveness is limited by the tool and your knowledge is limited to the tool's capabilities. It makes it hard to understand what is going on behind the scenes very well, making it nearly impossible for you to shift your perspective and find unique ways to apply your skills and find solutions.

The ability to take custodial knowledge ("memorized" or "taught" knowledge) and begin to understand it so well you see the wider applications and pivot...that has tremendous long-term value in my experience. You don't get that understanding by choosing a singular tool or a plug and play solution. You start getting that understanding by looking at the problem through multiple perspectives, the lens of multiple tools+multiple possible solutions.

That said, as a citizen data-scientist...absolutely, the ability to use these tools is valuable! And if a marketer told me they could use Tableau, Qlikview, IBM Watson, etc. - that is a great ADDITIONAL skill to have. But if that marketer tried to get hired as a data-scientist with marketing skills just based on their knowledge of those tools (rather than a marketer with some data analysis skills), I would be dubious.

I hope that makes sense, but this is just my general opinion about replacing technical knowledge with plug and play tools (in any industry). I am curious to hear what some of the data-scientists on here have to say :)

Thanks a ton. You made it so clear to me that I have good clarity now why should I invest good amount of time and effort for the ground level programming rather than being only GUI tool expert. Thanks for the detailed elaboration on this, really helps.

D. Shultz; your response was well worth a read, that principal holds true in almost any discipline or arena of knowledge and you explained it well. Good job!

Although not an official data scientist, I've been in the data business for a while.
Please don't give up on Python simply because a tool like Azure Machine Learning Studio has a drag-and-drop interface. You will limit yourself if you do. A lot of “data science” is data cleanup and feature engineering, much of which can require custom logic based on the task at hand. Coding will help you gain a better understanding of what's going on, as well as give you more ways to dive into analysis.

Tableau, Power BI, etc. are great tools to help tell your analysis story and allow business users to dive into the data. Power BI is cheaper and can produce simple dashboards relatively quickly. Tableau costs more but allows you (and your users) more visual data analysis. It’s true that Python & R can produce quick visualizations to help you understand the data. However, to go to the next step requires a lot more coding than it’s worth.

In conclusion, I would recommend that you keep up with your Python to scrub the data and enhance your analysis capabilities. Use a tool like Tableau or Power BI to present & share your results.

RSS

Videos

  • Add Videos
  • View All

© 2019   Data Science Central ®   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service