I have a question? I used this RE Agent,
https://lucidrealty.com/blog/ , he had some crazy crunched data (more than I felt necessary but awesome to see). I use apples Numbers for personal and Excel and SQL for work, so I know the limitation of big data on Numbers and Excel for the iPad. How are you able to deal with large data sets and crunching/parsing all of that on an iPad?
What do you mean by »large data sets«? Anyway: I just analysed a data set for a job application: 11 variables, about 10000 values per variable, over 2 periods. The thing was provided in an Excel file containing three work sheets. It was defined that, (for
various reasons) using Excel was not an option.
Disclaimer: the above is nothing I call „large data set“. 🤓
Task was to identify which variables and how these effect 2 specific ones from the set.
Due to various reasons, I had only an 2020 11“ iPP, iPadOS 16.4 with me (I managed to get a bluetooth keyboard and a mouse so 😃)
Without going to much into detail - I could elaborate on that, and I actually might put the whole story on my github - the requirements where basically to do descriptive and inferential statistics, t-tests, ANOVA, some external explorative statistical methods like heatmaps of correlation matrices, and machine learning. Additionally you had to write or implement around 60 functions/calculations and run them over this data and summarise these in a table (this table and relevant results and graphs from the explorative and statistical analysis had to be presented, all code documented). you had exactly 3 days.
I (almost) managed to do this in the given time
USING the iPad. 🤣🤪
I used
Carnets,
Juno,
Textastic,
Koder, R installed locally in
iSH (
apk add R
), R Studio via
posit.cloud,
Overleaf for the technical documentation (about 20 A4 pages of commented code and 10 pages of graphs), Apple’s Keynote to put together a (maximal) 10 slides presentation (that had to be presented via Microsoft Teams 🤕)
As I wrote, I
managed - I was able to "prototype" R and python code and run it to some extend locally...
but I had to generate local test data first, because a variety of modules, which makes your life easy when you want to analyse things, are simply not available in Python or Ron iPadOS… and you can’t simply compile them, due to the boundaries set by Apple.
At the end I solved most of the analytical part as well as the documentation via the mentioned online platforms.
To present in MS Teams on iPadOS, it turned out to better export the Keynote presentation to images or a PDF, “presenting“ in Teams means switching to the other app (Photos, Files/FileBrowser) but for example you do not longer see your video in Teams when you present (seriously? 🙄).
partly this feels like moving through honey with high viscosity or even non-newtonian fluids: if you push it, you hit something solid. (E.g. you can‘t simply run a calculation locally or in Safari in the background and switch to another split view of Keynote and Textastic/Koder/Files to import or copy data into it and work on your presentation. The calculation in the background might simply stop for a variety of reasons.) Be prepared to experience a work flow/procedure which oscillates between moments of pure delight and happiness followed by staggering disbelief resulting in hectic activity to search for, or create some form of workaround… if possible.
To all who get there stuff nicely done on iPadOS: Excellent!
To all who come up with something paraphrasing the Mandalorian creed “but you have to follow the iPadOS way!”:
you will get your Mandalorian helmet taking of your head (probably).
Accept the consequences. Go and take a bath in the living waters of ma
ndalcO
reS. 😁
IMHO there is importance in "how you do it" and not only in "can you do it".
Obviously YMMV.
TL;DR:
- you can analyse large data sets USING an iPad Pro in spring 2023.
- But only as long as you have access to online services.
- And only if you have a fast internet connection to transfer the data there in the first step.
- You can work happily in a browser window.
- Doing this on iPadOS is, if you are competent and lucky, at best a workaround in times of dire need.
- Be prepared to need possibly significantly more time then on macOS (obviously I am not referring to local calculation times on AS, but the whole shebang incl. uploads, reopening apps or re-establishing connections, etc.)
EDIT: I forgot to mention, obviously I tried to use Stage Manager to adress the multi tasking-related problems. The main problem here was, that just running it occupies a lot of RAM. Which then impacts on your ability to process the data locally, Safari disconnecting in the background, etc.. Not to mention the little problems or annoyances plaguing it. To be fair: RAM won’t probably be a problem on a 16 Gb RAM iPP - but in the meantime I have - just for fun - run the whole stuff locally on an older Macbook with less RAM than the iPP I used. Yes, the actual calculations are often slower than on the iPP. Significantly. But on MacOS the whole procedure not only feels faster, overall
it is faster done.