Technical help for data projects like this is expensive and, therefore, rationed to the point of near nonexistence. People learn to figure it out themselves. They magic things together with Excel, Access, and share drives. (1/?) https://bit.ly/30CRLQB 

And what people manage to cobble together with cardboard and masking tape is AMAZING and often does not fail in spectacular fashion. But it sometimes does! And you know what, the problem ISN'T Excel -- it's the lack best practices (testing, version control code review, etc) (2/?)

THAT'S where we can make a difference. Teach people that it's not enough to cobble together a shell script -- you have to share it someone so they can see what you're missing. You have to document your versions in a useful way, etc (3/?)

Don't dump on Excel users without acknowledging that they often HAVE NO ALTERNATIVE. "They should all learn R!" OK... with what time and money? It's like saying people living in poverty should try to make more money. (4/?)

My job has often been to unravel hacked together messes that people made, in the hopes that we can build them more reliable tools. It is my HONOR. I am inspired by what their love of their job pushes them to do. But you know what? Truth? (5/?)

Every Access database we replace with code has to be maintained by a software developer instead of the expert who needs it. And guess what? Software developers are usually MUCH MORE expensive. And budgets don't get bigger for each more elegant solution in production. (6/7)

So janky solutions, like it or not, are often the RIGHT call in certain conditions. So, maybe we consider Excel and Access magicians as PART OF OUR TEAM, instead of an abomination, and ask ourselves how we can support them? (7/7)