[pgviper]

Originally Posted by fecurtis

Yes exactly, in the real world no one is going to take the time to learn something new when they don't see the value in it. Especially since 99% of what they likely do can be done in Excel.



Because, unless you're working with a lot of data, Excel is suitable for 99% of most of what you'd see in the corporate world

Now if we're talking about scientific or statistical research, I'd imagine they're much more likely to know and use R since yes, Excel would be horrid to use at that. Hell I hate even using it for something as simple as regression analysis.

I know I haven't used it since college.

Agree to disagree then.

R has clear added value to a company over excel and the learning curve is at an all time low with the availability of dozens of free courses that teach basic data analysis and reporting.

On top of that, open source contributors such as Hadley continue to generate libraries containing high level functions that decrease complexity and increase functionality.

[Hawkeye]

Originally Posted by pgviper

Why do you say that? The primary draw back is that it has a higher learning curve than basic excel.
But at the same time, your time is much better spent learning R than diving into macros and creating solutions to problems that can be done in a few lines of R.
Those same lines can then be automated and built upon as opposed to excel.

I am going to assume you aren't employed by a company and just lay out some of the overarching reasons why it is impractical.

1. As stated in here, learning curve. Everyone knows Excel. Your upper level managers that like to dink with results, accounting, actuarial, and even IT all use Excel so it is not just a one department shift since everything gets handed off to someone.

2. Legacy work. You would not believe the number of legacy spreadsheets, models, processes, etc that are in Excel. To remake all of those is by itself a reason it is impractical.

3. Compatibility. I am less familiar with this but I would assume that there are not easy ways to export into all of the programs that are used by each division of a company.

4. Most important for last... control. There is this acronym that a lot of people hate, SOX. It pretty much means that you have to be able to control and validate all of the results you get. You cannot just rely on an open source piece of code that joe shmo put together to get you fancy results. THIS is the reason companies use more expensive, less capable programs.

4.5. Another reason people use more expensive and less capable programs is because of the support network. If there is a major problem running on a remote server or grid then there is an external party that you are already paying to fix it. This would be more convoluted with R (although possible and just as expensive probably if you just employed the right people in house)

[pgviper]

Originally Posted by Hawkeye

I am going to assume you aren't employed by a company and just lay out some of the overarching reasons why it is impractical.

1. As stated in here, learning curve. Everyone knows Excel. Your upper level managers that like to dink with results, accounting, actuarial, and even IT all use Excel so it is not just a one department shift since everything gets handed off to someone.

2. Legacy work. You would not believe the number of legacy spreadsheets, models, processes, etc that are in Excel. To remake all of those is by itself a reason it is impractical.

3. Compatibility. I am less familiar with this but I would assume that there are not easy ways to export into all of the programs that are used by each division of a company.

4. Most important for last... control. There is this acronym that a lot of people hate, SOX. It pretty much means that you have to be able to control and validate all of the results you get. You cannot just rely on an open source piece of code that joe shmo put together to get you fancy results. THIS is the reason companies use more expensive, less capable programs.

4.5. Another reason people use more expensive and less capable programs is because of the support network. If there is a major problem running on a remote server or grid then there is an external party that you are already paying to fix it. This would be more convoluted with R (although possible and just as expensive probably if you just employed the right people in house)

Assume away about my employment

1) Never stated for management to learn R, thats just silly to even think. There is a great feature called "EXPORT AS CSV/XLSX, etc..." which is the universal document for data.

2) Having business critical processes in the form of a spreadsheet just sounds insane to me. You have the same mindset that prevents progress, which is "ohh man, thats sounds difficult. Lets not do it."

3) See point 1

4) Not sure what SOX means but I can tell you this... plenty of the technology you use every day originated from an open source project.

4.5) You are right, but at the same time I have dealt with Microsoft support, enterprise and personal and have been disappointed with both.

[Hawkeye]

Originally Posted by pgviper

Assume away about my employment
This was not meant derogatorily but more as you are unfamiliar with some of the inefficiencies and complexities of a large business

1) Never stated for management to learn R, thats just silly to even think. There is a great feature called "EXPORT AS CSV/XLSX, etc..." which is the universal document for data.
So the solution to data presentation and delivery is to use Excel?

2) Having business critical processes in the form of a spreadsheet just sounds insane to me. You have the same mindset that prevents progress, which is "ohh man, thats sounds difficult. Lets not do it."
There are set budgets and most people are already working more than 40 hours as is, this is just not practical to think processes can be rebuilt from scratch in a new form without MAJOR investment (time and money) in the project. It most definitely would not be an efficient use of capital in the eyes of stakeholders since there are minimal realized returns.

3) See point 1

4) Not sure what SOX means but I can tell you this... plenty of the technology you use every day originated from an open source project.
https://www.sec.gov/answers/about-la...l.html#sox2002
http://www.investopedia.com/terms/s/...esoxleyact.asp

4.5) You are right, but at the same time I have dealt with Microsoft support, enterprise and personal and have been disappointed with both.
There is a finger to point at someone other than yourself if an auditor has an issue at least

Just some comments. I do think it is great in theory, just not practical in practice.

[pgviper]

Originally Posted by Hawkeye

Just some comments. I do think it is great in theory, just not practical in practice.

Valid points. I work in technology where the mindset is different but I can definitely see how companies in industries that progress at glacial paces may take some time to evolve.

Also, to your point 1: CSV is just a delimited file. It doesn't imply using Excel, it is simply an easy way to move around data in files.

[Hawkeye]

Originally Posted by pgviper

Also, to your point 1: CSV is just a delimited file. It doesn't imply using Excel, it is simply an easy way to move around data in files.

Yeah - I make pipe delimited files as deliverables all the time for my weird IT friends

[BMW F22]

Went back to 2010. Much faster now.