Let me provide some thoughts from my perspective.
I've recently transitioned from academia to industry and my title in the industry is "Data Scientist" in healthcare.
As everyone know, there is no unique way to define data science. Statisticians think they are data scientist while CS folks think they are.
I would not go into the debate.
I was trained as an "Applied Statistician" in a developing country back in late nineties. When we were undergrads, we had "pure stat" people who used to downplay us because we focused more on applications than on theory. For us (Applied Statisticians), now it is easier to align with the theme of data science than those "pure stat" group. Honestly, we were trained to use data to bring insights. We were trained not only to fit a model and know how it works but also to communicate in non-technical terms. This is what my understanding 20 years later. And I am not exaggerating.
Let me clarify one thing--data scientists do not necessarily design the databases. It is the data engineers who do that. Most large organizations have their separate team of engineers who develop and maintain the data-science platforms (DBs,Hadoop, etc.).
Do the statisticians need to know a bit of database? Yes, basic understanding is good enough. Most of the time all you will do is join some tables. For that you need to understand basics and you do not have to be a data architect for that. If you are working in a startup company, then you may be required to understand in greater depth, though.
How big of a deal knowing how to efficiently join tables? Not at all. Anyone with a decent knowledge about data can do it with some reading and practice.
In my work, I find statistics invaluable although most of the time we do not use many advanced techniques.
How media is portraying data science (such as deep learning) is what perhaps 1% of all the analytics an organization needs. Many large organizations do not hire people to do that. They purchase a solution for that as that is more cost-effective. For day to work where deep learning doesn't work, they need analytic people, aka statisticians.
In summary, the problems that data science try to solve are, most of the times, different than the problems that statisticians can solve. Being on both sides of the isle I can see how they can complement each other and how they both are relevant.
------------------------------
Enayetur Raheem, PhD
Data Scientist (HealthCare)
------------------------------
Original Message:
Sent: 08-27-2017 21:05
From: Steve Pierson
Subject: UPDATED: Your Thoughts on this Topic - "Data Science: The End of Statistics?"
Dear All,
Thank you for this discussion. I'd like to point out that some of these issues were addressed in last year's workshop of statistics and biostatistics department chairs: https://www.amstat.org/ASA/Meetings/Department-Chairs-Workshop.aspx. Dave Hunter shared the link to the new whitepaper last week: https://www.amstat.org/asa/files/pdfs/Chairsworkshop/WhitePaper.pdf.
Some of the videos may also be relevant to this discussion, particularly the ones by CMU Dean of Computer Science Andrew Moore, NCSU's Michael Rappa, and Winona State's Chris Malone. I highlight the last two because of their Master's and BA programs, respectively, on data science/analytics that were started from scratch. All the videos are linked form the first URL above. I found the Andrew Moore presentation particularly inspiring for their approach to bring many disciplines together to tackle challenging problems: https://www.youtube.com/watch?v=_aTBGDh8D78&index=7&list=PL9G4n1wtRTDTqwdSu8GhoqYIEIDkfHYMi.
Best,
Steve
------------------------------
Steve Pierson
Director of Science Policy
American Statistical Association
Original Message:
Sent: 08-24-2017 10:23
From: Gideon Bahn
Subject: UPDATED: Your Thoughts on this Topic - "Data Science: The End of Statistics?"
What Kelly and others provided for DS program are excellent, and a lot of people are working on it. As Michael and Andrew pointed out, it is not going to work if we are trying to make a jack of all trades after undergraduate. It may be better to specialize.
Yes, data scientists are in need for sure. But what are we expecting for the data scientists to do at work? Develop database, manage the data and analyze the data as needed (okay, each one of them needs to be defined)? We may have to clarity this first to start. Do we expect them to analyze the data and make a final report? I do not. I would rather to have them develop database, and manage the data, pulling some of data and cleaning them, for statisticians to work on them accordingly. Yet, if the DS understands some stats so that he knows what the statistician is asking data for, that will be great. Even within stats, we know how deep we have to go into based on the subject area, which will apply to a DS.
Hope this helps those who develop the program.
------------------------------
Gideon Bahn
Biostatistician at Hines VA Hospital
Research Assistant Professor at Loyola University Chicago
Original Message:
Sent: 08-24-2017 08:58
From: Stephanie Cano
Subject: UPDATED: Your Thoughts on this Topic - "Data Science: The End of Statistics?"
Well stated, Michael Mout!
Connect with us online! | Stephanie L. Cano, PhD Director of Statistical Consulting Analytic Focus, LLC 11467 Huebner Road, Suite 200 San Antonio, TX 78230
o: 210.641.2817 f: 210.641.9892 s.cano@analyticfocus.com www.AnalyticFocus.com |
CONFIDENTIALITY NOTICE: This e-mail, intended only for the addressee(s), is confidential and may contain certain personal and/or privileged information. Any dissemination or use of this e-mail by anyone other than an intended recipient is strictly prohibited. If you are not a named recipient, you are prohibited from any further viewing of the e-mail or any attachments or from making any use of the e-mail or attachments. If you believe you have received this e-mail in error, please notify the sender immediately and permanently delete the e-mail, any attachments, and all copies thereof from any drives or storage media and destroy any printouts of the e-mail or attachments. |
Original Message------
"Jack of all trades, master of none." Is the appropriate quote for Data Science. Data Science purports to cover Programming, Data Base development, and Statistical Analysis.
I have worked in all those fields in my 40+ years as a stats guy and it is the extremely rare person who can do all of those things well without having spent years working in each area. I have no problem with having a Data Science team with an experienced leader who designates experts in each of those three areas for the appropriate development.
Certainly, DS does not mean the end of Statistics.
Beware the Data Scientist who claims to be able to:
- develop complex DB's,
- write programs that can create that DB and generate reports from that DB AND
- do complex statistical analysis of that data.
DS is an appropriate field of study or for a survey course; however, if someone only gets an MS in DS then I suspect they will not be able to do any one of those three things very well. I think we can all agree that a Masters in Stats does not guarantee an expert analyst, it prepares in someone ready for an entry level position. If I were still working, I would never hire someone with a DS degree for an analyst position nor, I suspect, would a programming manager hire them for a programming position. Same for a Data Base manager.
I would recommend colleges have survey courses in DS, with appropriate deeper levels of study in one of the three specialties. Ar maybe something like a DS degree with required specialties in one of the three areas of study.
Mi
chael L. Mout, MS, Cstat, Csci
MIKS & Assoc. - Senior Consultant/Owner
4957 Gray Goose Ln, Ladson, SC 29456
804-314-5147(Mbl), 843-871-3039 (Home)