proc SQL

noetsi

Fortran must die
#21
I work in a dysfunctional environment. Its about power and empire building. IT does not exist here to serve the customer. They exist to own the data and fairly straightforward issues like running queries are intensely political issues. I got chewed out royally by my boss two weeks ago for admitting to IT I was unsure about a specific query (although this was true) because IT would use it as a wedge to take over part of our unit with senior leadership. IT owns the data, including knowing how to join the tables, and getting that information is a very painful process.

We can not even talk to the individual programmers (who build the data tables). We have to send formal requests through several layers of management.
 

bryangoodrich

Probably A Mammal
#22
I work in a dysfunctional environment. Its about power and empire building. IT does not exist here to serve the customer. They exist to own the data and fairly straightforward issues like running queries are intensely political issues. I got chewed out royally by my boss two weeks ago for admitting to IT I was unsure about a specific query (although this was true) because IT would use it as a wedge to take over part of our unit with senior leadership. IT owns the data, including knowing how to join the tables, and getting that information is a very painful process.

We can not even talk to the individual programmers (who build the data tables). We have to send formal requests through several layers of management.
That's when you tell your boss to kiss your ***. Be a trouble maker. Make waves. Rattle cages. That's how you make changes! Of course, I enjoy being that guy ;)
 

noetsi

Fortran must die
#23
That's when you tell your boss to kiss your ***. Be a trouble maker. Make waves. Rattle cages. That's how you make changes! Of course, I enjoy being that guy ;)
You are younger, likely have more savings, and are far more useful to your employer than I am. I can't afford to get fired.
 

bryangoodrich

Probably A Mammal
#24
You are younger, likely have more savings, and are far more useful to your employer than I am. I can't afford to get fired.
You need more confidence in yourself. Go apply to other, better jobs. If you land one, bam. You're in a better position. If not, you lose very little.
 

noetsi

Fortran must die
#25
lol. I would like to have a job where I actually get to analyze data, as compared to querying more. But those jobs are pretty rare outside academics.
 
#26
Yeah, I am surprised how much you all use SQL, etc. Data steps are always a weaker point for me, but I get to delve into analytics and write. The down fall is that it can be disheartening going through the publishing process sometimes and trying to care about repeatability.
 

bryangoodrich

Probably A Mammal
#27
lol. I would like to have a job where I actually get to analyze data, as compared to querying more. But those jobs are pretty rare outside academics.
There's plenty of jobs wanting people to analyze data, but the reality is that will always be a small part of the work load. A common number is that 80% of an analysts work is data munging, which is true more or less. Rarely does data come well defined and integrated that you can just take it as-is into an analysis. Even if I have pristine data and want to do a graph analysis, unless it's already loaded into a graph database, I still have to rearrange the data to be used for a graph analysis. That takes work outside of just analyzing data. Then even if you analyze data, you're probably gonna spend just as much time presenting it to show its value, make it useful. And that's really how you find those jobs. You're selling yourself as the person able to take their business questions, identify what decisions they need to make, and you'll give them the informational tools to make better decisions, to get data-driven answers. Couched in that way, every company needs analysts.
 

noetsi

Fortran must die
#29
I always spend most of my time "analyzing" data by doing queries and cleaning it up to analyze as discussed above. I expect that. What I meant is that I get asked to do analysis maybe 4 times a year period (as defined above). The rest of the time I am just pulling data for others to use.
 

bryangoodrich

Probably A Mammal
#30
You could always ask the people requesting data if they need more than just the data. Put on your business analyst hat, ask them why they need it, what are they trying to solve, drill down to the crux of their problem. Then show them the right analytical solution. Sell it. Then they'll want you to do the analysis for them, unless they think they can just do it better. Rarely have I seen anybody say "no thanks, I don't need more information" unless they're on a tight timeline and just need the data to report B, S, F, and U. That's if you want to go out there and find more interesting work to do. I've spent 5 years doing it. Unfortunately, my old manager drove that train, and it sucked most of the time. Now I can go out and find interesting problems to solve at a larger scale (finally!)
 

bryangoodrich

Probably A Mammal
#32
I bet most people asking for data don't know what to do with it!
Typically, no. When I'm not lazy and actually interested in taking on more work, I'd make sure to wear my BA hat and keep asking why, why, why? Until I get to the root of what they're trying to do. Usually clients have a vague, flowery notion of what they want. A data scientist's job is to unpack that into all the components of what it is they're really trying to do (much in the same way people asking Qs here don't really ask the right Qs half the time). That process of decomposing the problem lays out the work to be done in a clear operational way that you can then tackle it in a clear and objective way. Otherwise, it's "I need X" and then they come back next week "I need Y" because X wasn't what they really wanted and you end up providing little value as a data professional and end up wasting a lot of time.
 

noetsi

Fortran must die
#33
Well, in my experience anyway, they know why they want it. They may not understand what you produce or ask for the wrong data, but they know what the want. Its your role to get what they want not what they ask for if you can. And, this is the most difficult part, explain the limitations of the data and methods. Talking about uncertainty in either wins you few friends.

Of course, the fact that I have doubts about my understanding of statistics makes this harder for me. I got a Master's in Measurement and Statistics and spend lots of hours reading it (its why I came here in the first place) to address this and still am not too sure about that :p
 

bryangoodrich

Probably A Mammal
#34
If you think business people know what they want, then I can't imagine you've worked with them enough. They may *think* they know, but as soon as your scratch the surface and dig deeper you'll find a chasm of understanding. They rarely do the due diligence to really get detailed requirements around what they want. That's what a good business analyst does (and I'm scarcely even that good at it). Business users don't know the technical details, which is where BAs or business tech analysts specifically translate business requirements into those specific technical requirements, so the tech user can most easily get them what they want. But to think they just know what they want .. no ... not they don't. The one benefit of my old position was the ability to work directly with business stakeholders and play that BA role for analytic projects. Those typically don't have that BA intermediary to flesh out the details. It was up to me to do that in a rather agile way (wearing a PM hat to some degree)