Data Analysis Opportunities in Science Employment: Insights and Perspectives

[Monte_Carlo]

I've been visiting this section of the forum for some time and I'm curious about the opportunities in sciences for data-crunchers. I imagine many of the visitors here are also participating in other Internet forums and I'm eager to learn opinions about future of science employment in general. Some of the members voiced concerns over the prospects under current economic conditions but I think overall optimism prevails. Anybody has experience transitioning from being jobless to the state of gainful employment in science?

 

[Monte_Carlo]

A note on data crunching: I just did a mini-study using data collected from this forum. I looked for ratio of replies to views per post. As a sample, I thought to examine posts on first five pages. Interestingly, this variable (replies/views) is not normally distributed, in fact, the distribution appears to be negative exponential (see the attached histogram), with mean of 0.008. So, I conclude that on average a post needs to be viewed by 1000 individuals to get 8 responses. I'm going to continue my study, let me know your thoughts...

 

[fss]

Data crunching? Any "scientist" who only crunches numbers isn't a very good or useful scientist. A good scientist determines what the crunched numbers mean, designs the mechanism for obtaining the numbers to be crunched, and probably writes the code that does the crunching.

 

[Ben Espen]

If you are a true data lover, there are some employment options. I suggest you look into the http://analytics.ncsu.edu/?page_id=248" from North Carolina State University.

One should note that these graduates are not really scientists, as most people understand the term. This is a business degree, because the companies that need this kind of thing are often in finance, marketing, or advertising. As fss noted, most scientists and engineers do their own data-crunching.

 

[Monte_Carlo]

Thanks for the info, Ben. Whenever there is a new degree program like the one offered by NCSU, I'm curious how employable their graduates are. They show promising statistics.

What's your experience in data-mining and analysis?

 

[Ben Espen]

I am one of those engineers who do their own data crunching. I like stats, and I find the subject interesting, but it is an adjunct skill to process engineering and product design. I do DOE, SPC, and sampling plans pretty often, and I am also starting to look into code to automate some of these things that will generate reports on a regular basis.

 

[Monte_Carlo]

What language is the code in? What platforms do you guys use?

 

[Ben Espen]

There are a number of languages that my company uses, such as C#, that are suited to software development. What I do is use the scripting language built into our stats software (JMP) to formulate SQL queries to access databases, and then analyze the data that is retrieved as appropriate. This can range from Pareto charts to curve-fitting. This is all accessed through the GUI of JMP so that other people can use my work without having to know the details.

 

[Monte_Carlo]

Ben,

You mentioned that you're "...one of those engineers who do their own data crunching..."

How common is it, in your profession, to be intimately involved with mechanics of data analysis? The balance of subject-matter expertise and data proficiency seems to be off for "pure" data analysts. I understand from the way you worded it, that data churning is peripheral to your primary function, is that right?

Monte

 

[Ben Espen]

Monte, that is absolutely correct. My job is to get product out the door, and everything else is in service of that end. Every engineer does some data analysis, but the level of skill and interest can vary from using Excel to make a pie chart to custom written functions.