Pretty much all data heavy organisations use excel VERY heavily. And when nobody understands the model within them any more, they need retiring and are usually replaced with... Excel! This time with even more tabs and columns. To replace these things with computer models risks repeating the same problem the original sheet has: bus factors and complexities are hard, more so even in python/r than excel sadly. Maybe one day something will trump it, but that day is not today
SQL would like a little chat with you.
I do this for a living. I've spent basically my whole career (15 years full time professional at this stage) basically trying to kill excel. You can't, or at least I can't. You can add processes to it, you can programmatically read/write from it, but when it comes down to ditching it: every stakeholder is invested in excel. No other piece of office has the staying power that excel has, it will outlast us all
Because the tech-illiterate people who have authority only know "productivity" tools and couldn't care less about the opinions of the people who actually know what they're doing.
Those at the top are often more tech literate than I give them credit for. I suspect it is actually those armies of analysts that are holding it back
The Peter principle says otherwise, but analysts are a factor for sure.
I've witnessed the dilbert principle more
Yeah, I have access to database tools to do my job, but I don't know how to use those tools so I use Excel to do shit it really isn't optimized to do.
I am 100% part of the problem when I create a spreadsheet with formulas cross correlating data from 41000 entries, 9000 entries, and 1200 entries.
I disagree. It's way more that they aren't hiring the right people to do the job. I've been asked to do some analysis, but the only tool I know how to use is Excel so that's what I use to answer the mail. If I had access to a database person to help me build a better tool I'd be happy to not use Excel. But I don't so I do what I can to do my job.
It's surely a nightmare for long term usage but is there a software that can beat the functional reactive sort of auto updates when using spreadsheets with a few thousand rows of data? I'd have to actually use my brain to do the same thing as a pivot table in an array programming language.
Any sort of actual database will let you do it. SQL based the obvious answer, but they are all way harder to use than they should be. SQLite never got anything as good as excel sadly, and parquet still lacks a decent windows client. The WYSIWYG of excel really is so intuitive, nothing I know matches it
So Excel=cockroach?
I suspect slightly more useful than a cockroach. Believe it or not, it's actually good at what it does. That's why it's still here. And also why I'm in a job, as there are plenty of things it shouldn't be doing too
Naw other F1 teams just pay for custom data management software.
Supply chain management software exists. Do they meet f1 demands? Doubtful, but this is why you partner with a software company. They add more, you pay less, and give them some good sponsorship
Do they meet f1 demands?
If the alternative is a spreadsheet that gets updated manually, literally anything would be an improvement.
Markdown file on a shared windows drive enters the chat
You could use a database, and SQL. Or is that crazy talk?
Yes. Your boss needs to be able to double click on an email attachment otherwise it's like you never even did anything.
But in a language like Python there are established patterns for dealing with complexity unlike Excel lol
How do the handle version control? I'm yet to actually meet it in the wild
Version control for any kind of code can be done with git if that's what you are asking. There are other systems available too but that's the standard for most things now.
Python in excel is in git?? Better than I thought
Did you even read the article?
Yup. And extra extra, excel can be beaten for specific examples with lots of extra tooling. But you know what that tooling will also do? Generate excel reports
hmm generating static excel reports doesn't sound bad, as long as these are used just as output format - basically a familiar GUI.
Yeah, it is a good data dense gui, and is ubiquitous. But now it does python too! What a dreadful decision that was
Generating a report using Excel as a format is a lot different from using Excel as a database.
Pretty much all data heavy organisations use excel VERY heavily. And when nobody understands the model within them any more, they need retiring and are usually replaced with... Excel! This time with even more tabs and columns. To replace these things with computer models risks repeating the same problem the original sheet has: bus factors and complexities are hard, more so even in python/r than excel sadly. Maybe one day something will trump it, but that day is not today
SQL would like a little chat with you.
I do this for a living. I've spent basically my whole career (15 years full time professional at this stage) basically trying to kill excel. You can't, or at least I can't. You can add processes to it, you can programmatically read/write from it, but when it comes down to ditching it: every stakeholder is invested in excel. No other piece of office has the staying power that excel has, it will outlast us all
Because the tech-illiterate people who have authority only know "productivity" tools and couldn't care less about the opinions of the people who actually know what they're doing.
Those at the top are often more tech literate than I give them credit for. I suspect it is actually those armies of analysts that are holding it back
The Peter principle says otherwise, but analysts are a factor for sure.
I've witnessed the dilbert principle more
Yeah, I have access to database tools to do my job, but I don't know how to use those tools so I use Excel to do shit it really isn't optimized to do.
I am 100% part of the problem when I create a spreadsheet with formulas cross correlating data from 41000 entries, 9000 entries, and 1200 entries.
I disagree. It's way more that they aren't hiring the right people to do the job. I've been asked to do some analysis, but the only tool I know how to use is Excel so that's what I use to answer the mail. If I had access to a database person to help me build a better tool I'd be happy to not use Excel. But I don't so I do what I can to do my job.
It's surely a nightmare for long term usage but is there a software that can beat the functional reactive sort of auto updates when using spreadsheets with a few thousand rows of data? I'd have to actually use my brain to do the same thing as a pivot table in an array programming language.
Any sort of actual database will let you do it. SQL based the obvious answer, but they are all way harder to use than they should be. SQLite never got anything as good as excel sadly, and parquet still lacks a decent windows client. The WYSIWYG of excel really is so intuitive, nothing I know matches it
So Excel=cockroach?
I suspect slightly more useful than a cockroach. Believe it or not, it's actually good at what it does. That's why it's still here. And also why I'm in a job, as there are plenty of things it shouldn't be doing too
Naw other F1 teams just pay for custom data management software.
Supply chain management software exists. Do they meet f1 demands? Doubtful, but this is why you partner with a software company. They add more, you pay less, and give them some good sponsorship
If the alternative is a spreadsheet that gets updated manually, literally anything would be an improvement.
Markdown file on a shared windows drive enters the chat
You could use a database, and SQL. Or is that crazy talk?
Yes. Your boss needs to be able to double click on an email attachment otherwise it's like you never even did anything.
But in a language like Python there are established patterns for dealing with complexity unlike Excel lol
How do the handle version control? I'm yet to actually meet it in the wild
Version control for any kind of code can be done with git if that's what you are asking. There are other systems available too but that's the standard for most things now.
Python in excel is in git?? Better than I thought
Did you even read the article?
Yup. And extra extra, excel can be beaten for specific examples with lots of extra tooling. But you know what that tooling will also do? Generate excel reports
hmm generating static excel reports doesn't sound bad, as long as these are used just as output format - basically a familiar GUI.
Yeah, it is a good data dense gui, and is ubiquitous. But now it does python too! What a dreadful decision that was
Generating a report using Excel as a format is a lot different from using Excel as a database.