Sometimes, I work with bi-directional amplifiers and distributed antenna systems that are intended to improve cellular coverage inside of a building where there may be little or none of that.
I have a fairly expensive meter at my disposal to use for planning things like this, which analyzes different cellular carriers by frequency, and can output (messy, and with unescaped commas for notes, but eventually-fucking-usable) CSV files of the results -- with GPS coordinates of the measurement location.
This sounds amazing for a person like me in this line of work. But it is not amazing for a person like me in this line of work.
(As a preface for the rest of this, remember: This meter is a tool that is meant to be used in areas of limited or zero cellular coverage -- places where outside RF is problematic for whatever reason.)
1. The meter has a Bluetooth interface that connects to an app on a pocket computer. (This part works fine, usually, except the app often doesn't background properly and silently dies if the user uses their pocket computer to do some other task, which might be fine if the problem was ever reported. [Haha!])
2. The meter expects the pocket computer to have an Internet connection, so it can use that to upload its findings to The Clown. (This part often cannot work, because the whole fucking reason any of this is happening is because cellular coverage is shit inside of a random building.)
3. The meter expects that the pocket computer will provide GPS coordinates, even though it is intended to be able to be used indoors -- without network connectivity, or perhaps even in a Faraday cage. And while modern pocket computers are very good at providing some location data by various means as long as there is internet connectivity or GPS-esque data, all of them fail at this when there is neither Internet nor GPS available. It produces an error [Haha!] when there is no location information available.
4. It does not provide useful errors. It provides errors, but they aren't specific at all and do not promote productive troubleshooting or workflow. ("Oh, there was a problem with your measurement! [Haha!]" is the singular error.)
5. Sometimes, it will even produce an error [Haha!] but record the measurement anyway -- and without recording the error.
6. It stores nothing locally. When an error happens [Haha! Good luck!], it is impossible to quickly see if anything was stored at all, so the only clear path is to repeat measurements that result in an error [Haha!]. This often results in redundant measurements being actually-recorded, but who would know that at the time of measurement. (These measurements often take about 4 minutes each, so these errors [Haha!] and repeated measurements can consume significant portions of an expensive workday.)
7. (Your main point): Exporting a CSV file of [whatever-the-hell was collected] is possible, as long as I want to send it to Google Drive or some other Clown-based service. The CSV is only a few tiny kilobytes at very most, but it won't let me copy the CSV to my pocket-computer's clipboard, or send it in an email, or save it locally on the pocket computer, or send it with Bluetooth to my laptop. It has to be exported to a Clown-based service, and then it can be read from that Clown-based service by some other device. There are no other options presented, unlike in so many other apps in my pocket computer.
8. Continued: While the maker of this meter device has their own Clown, and this Clown is clearly extent on the Internet, this Clown is completely inaccessible outside of their pocket-computer app. I cannot bypass Step 7 by any official means no matter how deep my desktop computing prowess may be.
It is completely shit, and it appears to be the best thing available on the market in this space. (And it isn't even Chinese shit: The company that produces this meter is in Utah.)