Party Buzz Kill: modifying data
So Steve (SQL), Marsha (C), Bob (Python), and I (R) are at this party. We have TOTALLY cleared the room, especially now that Steve and I are deep into a debate about saving native data objects to disk versus storing data in a database.
I see my friend Monica enter from the kitchen, carrying a bowl full of punch. It's an awkward task and the fruity, sticky liquid is sloshing on the floor. Monica does data science, so I'm hoping she'll come to my assist. Sure enough, she places the punch bowl on the table and joins us.
She listens for a minute, then interrupts the pointless debate between Steve and I. "People who are math aficionados" she says, "are a lot more comfortable generating datasets on-the-fly. People like me enjoy relying on the safety and reliability of importing a structured dataset we checked earlier!"
Steve is happy to hear someone is on his side. Steve thinks I'm a knucklehead. There are many people who agree.
But Monica isn't done. "But you are correct - the question is technically two sides of the same coin."
"Sure, but there are advantages to not messing around with unnecessary overhead," I say. "Let's play with an example."
Read more about this exciting installment...
#rstats #sqlite #punchbowl
LinkedIn [in]structor | Data Science Consulting
1yI like generating my own datasets from public data sources. I will say, though, that it can become time-consuming to do this every time. So, I often stick to a dataset (especially ones that update to the latest data like APIs and FTP folders) until I get tired of it and want to try something new...
Epidemiology & Biostatistics Consultant a/k/a Data Scientist | Exclusive and innovative solutions for data science challenges in public health, research and education
1yHey everyone reading this - it's worth it to go to the blog post, because there is a function there that will knock your socks off!!! Thank you for making this argument actually interesting, Mark Niemann-Ross!