Even just in the context of doing data cleaning in R or Python, most data scientists aren't writing code that's readable, testable, documented, and part of processes that are replicable. Even if you're just cleaning data, you can do it well.
Except industry data doesn't look like an IPUMS extract.
Even just in the context of doing data cleaning in R or Python, most data scientists aren't writing code that's readable, testable, documented, and part of processes that are replicable. Even if you're just cleaning data, you can do it well.Except industry data doesn't look like an IPUMS extract.
The worse your data is, the more important that you be following good practices as far as coding for anyone to be able to understand or build on your code, or even for you to understand what you did and why when you come back to it.
Even just in the context of doing data cleaning in R or Python, most data scientists aren't writing code that's readable, testable, documented, and part of processes that are replicable. Even if you're just cleaning data, you can do it well.Except industry data doesn't look like an IPUMS extract.
The worse your data is, the more important that you be following good practices as far as coding for anyone to be able to understand or build on your code, or even for you to understand what you did and why when you come back to it.
No doubt. I wanted to emphasize that data cleaning of your typical econ data set is child's play compared to industry data.
Even just in the context of doing data cleaning in R or Python, most data scientists aren't writing code that's readable, testable, documented, and part of processes that are replicable. Even if you're just cleaning data, you can do it well.
Except industry data doesn't look like an IPUMS extract.
The worse your data is, the more important that you be following good practices as far as coding for anyone to be able to understand or build on your code, or even for you to understand what you did and why when you come back to it.
No doubt. I wanted to emphasize that data cleaning of your typical econ data set is child's play compared to industry data.
This. Industry data is about as incomplete and untrustworthy as it gets.
Most DS jobs just need SQL and a bit of R. Don't overthink it.Most don't even need a bit of R.
Yeah, I'm an economist with a Github full of R, and finding a job that lists R is like needle in a haystack compared to python/SQL. If you know R well, but nothing else, don't expect to be rolling in job opportunities.
The work is done poorly in industry. Low standards.
Even just in the context of doing data cleaning in R or Python, most data scientists aren't writing code that's readable, testable, documented, and part of processes that are replicable. Even if you're just cleaning data, you can do it well.
Except industry data doesn't look like an IPUMS extract.
The worse your data is, the more important that you be following good practices as far as coding for anyone to be able to understand or build on your code, or even for you to understand what you did and why when you come back to it.
No doubt. I wanted to emphasize that data cleaning of your typical econ data set is child's play compared to industry data.
The thing with industry people is that the standards are so low and the delusion is so high. There is nobody telling you that your work is cheet. As long as your manager is a doo/fus, you can literally do anything.
^Orrrr that industry data comes from the real messy world and isn't purchased for you nice and clean on wrds
The thing with industry people is that the standards are so low and the delusion is so high. There is nobody telling you that your work is cheet. As long as your manager is a doo/fus, you can literally do anything.
^Orrrr that industry data comes from the real messy world and isn't purchased for you nice and clean on wrds
Free rider problem notwithstanding, there's obviously the ultimate consequence of the entire data science team being axed if the work is poor.
that only strengthens my argument. they were doing bad work and it blew up.
The thing with industry people is that the standards are so low and the delusion is so high. There is nobody telling you that your work is cheet. As long as your manager is a doo/fus, you can literally do anything.^Orrrr that industry data comes from the real messy world and isn't purchased for you nice and clean on wrds
Free rider problem notwithstanding, there's obviously the ultimate consequence of the entire data science team being axed if the work is poor.
https://www.dailymail.co.uk/news/article-10158485/Zillow-plans-lay-25-staff-putting-end-home-flipping-business.html
You are a delusional. SQL is very easy.
Fortune 500 data scientist here. Basic SQL queries and basic R/Python data manipulation is literally all we do.Still miles above what your average applied micro monkey could pull of
What percentage of economists would pass the SQL portion of a coding interview?
Free rider problem notwithstanding, there's obviously the ultimate consequence of the entire data science team being axed if the work is poor.
https://www.dailymail.co.uk/news/article-10158485/Zillow-plans-lay-25-staff-putting-end-home-flipping-business.html
The data science team was not the problem with Zillow's home-flipping business. There is a fundamentally large adverse selection problem with the business concept. On top of that, the higher-level management saw the business making lots of money as the market was going up and pushed staff to raise their valuations, which is what caused the blow-up.
the standards are only low for DEI job candidates and the sons and daughters of important people who need to be given cushy jobs
which in practice are most data scientists it seems like
The thing with industry people is that the standards are so low and the delusion is so high. There is nobody telling you that your work is cheet. As long as your manager is a doo/fus, you can literally do anything.
^Orrrr that industry data comes from the real messy world and isn't purchased for you nice and clean on wrds
Free rider problem notwithstanding, there's obviously the ultimate consequence of the entire data science team being axed if the work is poor.
https://www.dailymail.co.uk/news/article-10158485/Zillow-plans-lay-25-staff-putting-end-home-flipping-business.htmlThe data science team was not the problem with Zillow's home-flipping business. There is a fundamentally large adverse selection problem with the business concept. On top of that, the higher-level management saw the business making lots of money as the market was going up and pushed staff to raise their valuations, which is what caused the blow-up.
Poor economic modelling might have played a role, too. The point was shoddy data science work doesn't go unpunished.
But it did go unpunished, in fact it was encouraged. lol.
Free rider problem notwithstanding, there's obviously the ultimate consequence of the entire data science team being axed if the work is poor.
https://www.dailymail.co.uk/news/article-10158485/Zillow-plans-lay-25-staff-putting-end-home-flipping-business.htmlThe data science team was not the problem with Zillow's home-flipping business. There is a fundamentally large adverse selection problem with the business concept. On top of that, the higher-level management saw the business making lots of money as the market was going up and pushed staff to raise their valuations, which is what caused the blow-up.
Poor economic modelling might have played a role, too. The point was shoddy data science work doesn't go unpunished.