You can now wield Python analytics inside Excel—but Microsoft says it’s safe
- Thread starter JournalBot
- Start date
Not gonna lie, being able to execute Python code within Excel sounds like a killer feature. Though I worry about the security holes it might open up.
Upvote
86
(87
/
-1)
Excel has needed an improved programming language for a long time, and I've always thought python was the answer, but why require running this in the cloud?
Upvote
106
(107
/
-1)
Well, default-settings matplotlib graphs aren't great to look at, but they're a heck of a lot better than default-settings Excel graphs for things like scatter plots and line plots, so I guess that's a step in the right direction.
Upvote
23
(24
/
-1)
That may be true, but the official logo for the Python programming language sure looks like a pair of Burmese Pythons, and not a British comedy troupe.
Edit: typo correction
Last edited:
Upvote
40
(41
/
-1)
Like, what or who is this feature going to be killing, and should we be worried?
Upvote
10
(14
/
-4)
Its going to kill my productivity, most likely.Like, what or who is this feature going to be killing, and should we be worried?
Upvote
24
(26
/
-2)
Why start something like this in the desktop app at all? Feels like they’re grasping at straws here. Just use Jupyter Notebook as an analyst, as MS, keep your new development in Azure or what are you even doing?
Upvote
-3
(10
/
-13)
PatientZero
Ars Centurion
Monetization (paywalls), data-collection (further profit on selling anonymized data in non-GDPR counties), etc.
Upvote
45
(47
/
-2)
Its going to kill my patience with end users wondering what does this button do, and oh call support because now it looks different.Like, what or who is this feature going to be killing, and should we be worried?
Upvote
20
(21
/
-1)
Microsoft has been using your ownership of one product to force you to buy another product for nearly 50 years. Why stop now?
Upvote
36
(36
/
0)
NOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
When it comes to being able to pass around research data, there is no bigger cancer than Microsoft Excel. It does things without telling you and then leaves it up to someone far down the line to clean up because nobody noticed until it was too late. See Excel genes.
When it comes to being able to pass around research data, there is no bigger cancer than Microsoft Excel. It does things without telling you and then leaves it up to someone far down the line to clean up because nobody noticed until it was too late. See Excel genes.
Upvote
81
(91
/
-10)
Because they're slowly trying to move the whole of Excel into the cloud. What you can do there is already quite impressive but you can't do VBA, and many businesses still make heavy use of that. Providing this Python functionality and making it run in the cloud is another step in that direction on top of the basic scripting they already have there.
Upvote
27
(27
/
0)
Upvote
2
(6
/
-4)
If you're using Python anyways, Plotly can produce some very nice charts and graphs. Can read a pandas dataframe too.
Upvote
15
(15
/
0)
. Beside the fact that Microsoft would love to host everything including your Windows OS boot from the cloud, hooking Excel into a Python backend in the cloud that they control is going to be so much easier than trying to actually integrate Python into the application itself.
Upvote
19
(19
/
0)
Many vendors including those of scientific and industrial technical packages provide scripting languages built into their apps to automate work flows etc. These scripting languages always seem to be very primitive and restricted. It would be great if they would commit to providing libraries that would allow full use of standardized general languages like Python.
Upvote
11
(11
/
0)
Actually many people did realize that. It was impossible to fix because MS wouldn't change it. Actually one of the annoying aspect of spreadsheets of format guessing.
Just use something like pyspread, a non-traditional spreadsheet application that is based on and written in the programming language Python.
Upvote
20
(21
/
-1)
That’s pretty exciting. I think a lot of people like me don’t have access to analytics and visualization tools nor the expertise in using them even if we did have them.
Upvote
0
(1
/
-1)
I was initially excited about this news. In spite of my efforts otherwise, the place I work runs on Excel. And dropping VBA for Python could make my life easier. But now I see here, all it looks like it's really doing is offering a way to create custom formulas? I don't know about other places, but in my particular work bubble that's a huge shrug.
On top of that though, to run it in a cloud container? And then say it's a security advantage? I don't often feel so actively insulted by marketing, but here we are.
On top of that though, to run it in a cloud container? And then say it's a security advantage? I don't often feel so actively insulted by marketing, but here we are.
Upvote
19
(22
/
-3)
Retired data scientist, and this is the most cursed thing ever.
Like, you have to get your organization out of Excel except for ad-hoc, non-operational stuff, and this just screams for your boss to demand you move your tools into Excel so they don't have to learn anything new. This is going to break your shit so bad down the road.
Like, you have to get your organization out of Excel except for ad-hoc, non-operational stuff, and this just screams for your boss to demand you move your tools into Excel so they don't have to learn anything new. This is going to break your shit so bad down the road.
Last edited:
Upvote
40
(41
/
-1)
I'm not sure it's fair to claim it not a security advantage, but it's definitely not as big of an advantage as they want to claim given recent news stories.
Most IT groups want to protect their end point devices. If I run my code on a container in the cloud it's not running on my computer. That means my endpoint is safe(er? I'm unwilling to claim a cleaver red teamer/hacker can't figure something out). You definitely need to worry about data leaks from Microsoft, but you don't have to worry about "Billy from accounting@bdgfkc 123.ru sent me this excel file better run it and see he needed me to see". Because then it runs in the container and successfully gets all the information sent to it from the excel file provided.
In the Billy from accounting situation you have to worry about someone outside your org a foothold to pivot in your org and possibly getting access to all your information. In the current situation you have to worry about someone getting access to information stored on MS servers. Not good, but not as bad I'd think.
Upvote
16
(16
/
0)
PatientZero
Ars Centurion
Interesting... and profitable.
"We see you're having trouble viewing this Excel file because it was created with Microsoft
Python in the Cloud -- would you like to upgrade one-time for $5.99 to view this file?"It's basically Business Software Microtransactions.
I already hate it.
Upvote
18
(20
/
-2)
My biggest concern right here. Quoted from the verge article on this: "Python calculations run in the Microsoft Cloud, with the results returned into an Excel worksheet."
Upvote
3
(3
/
0)
Python + Pandas.
Or R.
Fk code that you can't read in whole, but have to look at cell by cell.
Or R.
Fk code that you can't read in whole, but have to look at cell by cell.
Upvote
9
(9
/
0)
So, the main problems with anything operational in Excel has nothing to do with the coding language. It has to do with the inability to build the tool from code, to version it/diff, to document it in a competent way, to get logs, and so on, and all of those problems remain here.
Upvote
36
(36
/
0)
Matlab has the same endpoint problems that excel does. It's really useful, but if you need to do anything that the tool doesn't handle, your API options and ability to run it as a local service are shit. Plus you have the same closed-box problem with code management.
Upvote
11
(11
/
0)
Seconded. I dabbled with it a few years back. At the time there was even a Plotly offline library that still had pretty good functionality. Though, I haven't touched it since and don't know if they've neglected it to push some cloud service. Not that I wanted to let it go, but we started getting more secret squirrel stuff and The Man still frowns on OSS.
Upvote
2
(2
/
0)
Plus you have to pay for it. Plus more for the various toolboxes. Plus more if you want to scale the workload up to a cluster.
Plus the machine you run it on needs to be able to talk to a license server (for most sites which do not have unlimited site licenses). So, if you're working from home, and your internet happens to cut out, you're done working on Matlab. (That's why I switched to Python in grad school, way back when numarray was still a thing.)
Last edited:
Upvote
10
(10
/
0)
I'd settle for just being able to run SQL SELECT and UPDATE statements against the Excel sheets as if they were tables.
It's pretty trivial to do via ODBC against the Excel file, but it'd be way more handy if it were just inside Excel.
Upvote
12
(12
/
0)