Data Science how important is statistics


#1

Hi All,

I am totally new to the field of data science. I read about basics.

I am confused need some expert advice on help me understand the difference.

I so far understood DS is combination of programming skills (R, PYtthon ,SAS … etc) and statistics. The courses I checked so far are focusing on 80% programming skills and 20% on statistics.

Since i am totally new and have no idea of statistics too… I thought if I pick a course which focuses on 80% statistics and 20% programming that might help.
My understanding is programming is a mode of implementing statistics.

Please correct if my understanding is wrong and how to figure out the right fit for me. I want to have strong foundation in basics.

Thanks in Advance.
Siri


#2

Hey Sirich,

You should learn statistics before focusing on coding/tools. You can spend a little time learning tools, but your primary focus should be stats since tools are of no use if you are not able to read insight(s) from it. Using a wrong method or tool can be extremely dangerous as well.

I hope this helps.


#3

Unless you are going to get into say a core field where statistical interpretation plays more significance than quick win tool based solution, you can learn on the go. There is no pressing needs to pressure yourself to suddenly become a master in statistics.

You definitely need to brush up some basics but 90% of current so called data scientists are good with tools, ability to apply techniques and create consumable results but have no clue about the working or science behind those techniques. And IMHO you dont have to get to the science behind these techniques as the consumer hardly has any clue even if you talk the technique language let alone the science behind the technique language.

Put it simple, say a retail customer wants some magic out of data, he wants to see the outcome and wont care how you arrived at it. If the retail customer is also a statistical major he may poke you around model related questions to see if you tried different methods. And now if you are a statistic major your conversation will bring joy to you but the outcome is still going to be the same where he will already know what to take out of your output, request for a different technique and interpret on his own.

So learn statistics, but more so learn about techniques in detail, apply for data set and continue to learn how you can interpret and learn about different techniques and models so you will learn to say no to a model for a specific type of data.