"Data Scientist" is such a weird title for somebody who is basically a statistician with good IT skills. I think in some market verticals they're simply called "Quantitative Analysts" and make very good money.
In a previous job I did lots of analysis on very large data (at the time) data volumes, millions of structured or unstructured records, homo and heterogeneous datasets. Lots of aggregation, sifting, sorting, simplifying, deduping, summarizing, etc. All in support of similar kinds of things that "Data Scientist" positions seem to be intended to support. But the output was not a statistical model, or a machine learning exercise or some other similar. It was the distillation of gigabytes of data into a handful of slides and a report. Usually with a virtuous cycle of feedback directly into software development to improve and expand the next go-around.
But almost no statistics. Very very little, and what I did was very basic stuff.
What is that kind of job called? In my day we called it a "Data Analyst" but I don't see that around much.
I'm going to make a prediction, "Data Scientist" as "Senior Statistician" is going to be short-lived. I don't think they're going to provide the value companies think they will in most cases. "Data Analyst" is much more general purpose and useful cross-domains, except most Data Analyst don't have proper statistical training.
A Data Analyst with statistical training would be a much more useful tool to an organization seeking to make sense out of large volumes of data than a Senior Statistician as they'll have a much wider variety of tools at their disposal than just looking at the world through the statistics lens.
Bonus, jobs advertising "Data Analyst" can demand things like machine learning AND entity extraction AND automatic summarization AND data sanitation AND automatic correlation analysis AND automatic colocation analysis etc.
Most of the jobs I've seen looking for Data Scientists are for companies that are probably going to try and end up using them as high-priced Data Analysts, except the job reqs are all wrong and the candidates that get hired are way over qualified.
But this role is still evolving I suppose, IBM [1] views it as an evolution from the business/data analyst. So they definitely seem to be on the side of not so much statistics and more analysis.
I'm going to throw it out there that a Statistician with poor IT skills nowadays is like a carpenter with poor measuring, cutting and hammering skills.
In a previous job I did lots of analysis on very large data (at the time) data volumes, millions of structured or unstructured records, homo and heterogeneous datasets. Lots of aggregation, sifting, sorting, simplifying, deduping, summarizing, etc. All in support of similar kinds of things that "Data Scientist" positions seem to be intended to support. But the output was not a statistical model, or a machine learning exercise or some other similar. It was the distillation of gigabytes of data into a handful of slides and a report. Usually with a virtuous cycle of feedback directly into software development to improve and expand the next go-around.
But almost no statistics. Very very little, and what I did was very basic stuff.
What is that kind of job called? In my day we called it a "Data Analyst" but I don't see that around much.
I'm going to make a prediction, "Data Scientist" as "Senior Statistician" is going to be short-lived. I don't think they're going to provide the value companies think they will in most cases. "Data Analyst" is much more general purpose and useful cross-domains, except most Data Analyst don't have proper statistical training.
A Data Analyst with statistical training would be a much more useful tool to an organization seeking to make sense out of large volumes of data than a Senior Statistician as they'll have a much wider variety of tools at their disposal than just looking at the world through the statistics lens.
Bonus, jobs advertising "Data Analyst" can demand things like machine learning AND entity extraction AND automatic summarization AND data sanitation AND automatic correlation analysis AND automatic colocation analysis etc.
Most of the jobs I've seen looking for Data Scientists are for companies that are probably going to try and end up using them as high-priced Data Analysts, except the job reqs are all wrong and the candidates that get hired are way over qualified.
But this role is still evolving I suppose, IBM [1] views it as an evolution from the business/data analyst. So they definitely seem to be on the side of not so much statistics and more analysis.
1 - http://www-01.ibm.com/software/data/infosphere/data-scientis...