I read a lot of stuff and respond to a lot of it. Stuff I don't read I open anyway to make th icon go greyscale so it doesn't mock me with th colorness

Pumpkin King @ 31/10/2017 18:14

anyone else ever see a wall of text quotes from vlady/cat and just scroll by? because that's what I do. every time.

Just did

Grandma just found out there's an Uzbekistan. She had no idea how to say it. She asked me about it and when I said the name she started at me blankly for almost 30 second before asking "U-Pakistan??"

Her mind is blown

drrrrr @ 31/10/2017 20:34

Grandma just found out there's an Uzbekistan. She had no idea how to say it. She asked me about it and when I said the name she started at me blankly for almost 30 second before asking "U-Pakistan??" Her mind is blown

There is an elderly lady at my complex who called the office yesterday because she "feared for her safety". I had to assure her that everyone who lives here has gone through our screening process and has passed.

"There's a Chinese girl who came out of the apartment. She said hi. She seems nice. And the Mexicans next door are okay. I think there is a guy above me...I think he is a black man."

I think she's scared because the only white person in her building just got evicted yesterday and had to move out, so she's the last remaining white member.

Every new Rose story I think of as new material. lol

Poll of the Day #127

Do you say 'Pakistan' or 'P-A-Ki-S-Tan'?

drrrrr @ 31/10/2017 20:34

Grandma just found out there's an Uzbekistan. She had no idea how to say it. She asked me about it and when I said the name she started at me blankly for almost 30 second before asking "U-Pakistan??" Her mind is blown

I had to help one of my Kentucky family members pronounce it today. She didn’t know terrorists came from countries other than those in Trump’s immigration ban...

Really?

remember when america and pakistan pretended to be allies for most of the cold war even though we've always kinda hated each other

sir ganache bbq:spin: @ 31/10/2017 22:43

Poll of the Day #127 Do you say 'Pakistan' or 'P-A-Ki-S-Tan'?

Yeah, Pakistan has always annoyed me for messing with the istan meaning. Plus taking the end of one province for the "stan" because PAKSB didn't sound cool.

index/match >>>> vlookup

The snowmen are in Hollywood

vladykins @ 1/11/2017 12:00

sir ganache bbq:spin: @ 31/10/2017 22:43 Poll of the Day #127 Do you say 'Pakistan' or 'P-A-Ki-S-Tan'? Yeah, Pakistan has always annoyed me for messing with the istan meaning. Plus taking the end of one province for the "stan" because PAKSB didn't sound cool.

or even PNKSB, ha

but yeah, they were just shoehorning a backronym for pak-istan (meaning 'pure land') with the anaptyxic i due to the phonotactics of the suffix following a closed syllable.

we're spanish bbq:spin: @ 31/10/2017 16:11

vladykins @ 31/10/2017 15:46 SQL mostly for just our web surveys, we crunch a lot of crap using Python (like when we have to pull in data and output into a bunch of readable tables) or we crunch it in SAS). We've been using Tableau for a lot of data visualization, though we're working with Power BI more now because we still have idiots who want their outputs in Powerpoints and Tableau is a pain for that. We have been thinking of integrating out web survey crap all on SPSS so we don't have to have separate data collection and analysis systems, but the firm is slow as hell on shit like that. good stuff, always interested to see what people are using in the field. any particular flavor of SQL? i was a postgres fan for a while, but now amazon aurora just solves so many other problems. in fact, i've even been moving certain NoSQL stuff to aurora (especially for production environments). mongo is still easier for cranking out prototypes, but most of what i work with is very easy to port between relational and nonrelational environments, so why not. for python, are you pulling in data using stuff like requests and beautifulsoup? or is it just reading in csv files and munging with pandas? any numpy/scipy stuff? for some dataviz, seaborn is easy and bokeh is cool, but few firms want to pay analysts to code dataviz -- so i've heard of other firms leaning to powerBI for the same reasons you just mentioned, which makes me wonder where tableau is going to sit in the future. they seem to be veering towards boutique solutions sold by consultants rather than office software you can slowly get going out of the box. i've barely touched SAS or SPSS. licenses are too pricey and if i really want to do stats, R is free. next question: when your firm is hiring (not that i'm looking for another job), what sort of software skills are they looking for? or is the emphasis just on analytical thinking with the assumption that software skills can be learned on the job?

I'm dealing direct with SQL server in a lot of cases (either playing in Management Studio or pulls from SAS, Access, or Excel), since we tend to pull down the tables as flat files and then process the data, rather than doing it live online. Mainly we're only playing with SQL on our legacy survey pages where we need more flexibility than we have from our newer out-of-the-box web survey tools. It doesn't help that we have to access control shit like crazy, which means trying to run anything new means infosec comes in and shuts down half the features that make the new stuff worthwhile (and why our legacy software still floats around 15 years later). OTOH, our legacy shit still works on any browser, so I can't complain too much.

Pyhton we're mostly just using it to chop up csv files to do some data transformation to ultimately analyze in Excel or Power BI/Tableau. Previously we did all the data transformation and derived variable work in SAS, but then you have to read in and create a bunch of datasets, then output into excel in certain places (which prior to SAS Access in version 9 we had to use a third party software called DBMS Engines to help with that). Python is so much more efficient, so we've been exploring how we can better use it.

Our particular group is 30-35 economists, statisticians, and mathematicians (usually applied math). Our staff positions are usually hired right from undergrad, they work for 2-3 years, then go off to grad school. In their case, we look for folks who seem to be able to learn to program, rather than programming experience (as noted above, most haven't played much with Excel at that point, so we start them easy with vlookups before we get more complicated). Any direct hired seniors and above tend to have an advanced degree and often have much more programming background. Still for our web survey work, we've been thinking of hiring someone with more of a data sciences background to help push us forward, since we have the survey methodology and operations research side covered but could use more help on the tech end.

vladykins @ 31/10/2017 17:45

sir ganache bbq:spin: @ 31/10/2017 16:11 vladykins @ 31/10/2017 15:46 SQL mostly for just our web surveys, we crunch a lot of crap using Python (like when we have to pull in data and output into a bunch of readable tables) or we crunch it in SAS). We've been using Tableau for a lot of data visualization, though we're working with Power BI more now because we still have idiots who want their outputs in Powerpoints and Tableau is a pain for that. We have been thinking of integrating out web survey crap all on SPSS so we don't have to have separate data collection and analysis systems, but the firm is slow as hell on shit like that. good stuff, always interested to see what people are using in the field. any particular flavor of SQL? i was a postgres fan for a while, but now amazon aurora just solves so many other problems. in fact, i've even been moving certain NoSQL stuff to aurora (especially for production environments). mongo is still easier for cranking out prototypes, but most of what i work with is very easy to port between relational and nonrelational environments, so why not. for python, are you pulling in data using stuff like requests and beautifulsoup? or is it just reading in csv files and munging with pandas? any numpy/scipy stuff? for some dataviz, seaborn is easy and bokeh is cool, but few firms want to pay analysts to code dataviz -- so i've heard of other firms leaning to powerBI for the same reasons you just mentioned, which makes me wonder where tableau is going to sit in the future. they seem to be veering towards boutique solutions sold by consultants rather than office software you can slowly get going out of the box. i've barely touched SAS or SPSS. licenses are too pricey and if i really want to do stats, R is free. next question: when your firm is hiring (not that i'm looking for another job), what sort of software skills are they looking for? or is the emphasis just on analytical thinking with the assumption that software skills can be learned on the job? I'll post on this tomorrow, since I'm outside on phone while giving out candy, but we are looking for more of a data science guy in the near future for our survey team.

data science meaning more towards the data engineering and cloud dev side of life? or more towards the realm of statistical learning?