Data Studio-Friendly Export

Hi! I’d really like to be able to import my data into a health report I have created in Google Data Studio, but the Google Sheet export is not friendly for that, as each entry has its own set of headers, and a lot of specific event information that differs from session to session.

I’m looking for an export that could include all of the data that is in the current export, but without the details of the events. This way, you could have one single row with a heading on it, and a row of aggregated data for each day/sleep session.

What I’m envisioning looks a lot like the chart at the bottom of the SleepCloud page.

Thanks in advance!

Hello Lisa,

we have many thanks we plan to get rid of the repetitive headers… and we will see what we can do about the other info which isn’t easy to get from the sheet…

BTW I looked at the recordings you shared…

This is a device specific issue. Vendors tent to enhance the audio experience apply filters to speakers or mic which break sonar. So that the recorded sonar signal gets into the hearable spectrum…

You can try to play with the various sonar settings when you do sensor test… there is F18-22 (Default) but also F18, F19, F20 and try to set signal strength a bit lower. Maybe you can find a combination of settings which works well on your phone…

Dear @Lisa_Clarke, we also have a conversion tool that you can use to convert the export CSV from Sleep as Android to a JSON that is better structured for further use.
Take a look!

Why not export in a format that is ready for use instead?
The current format turns what should be a simple as opening a spreadsheet into hell.

(edited: been trying to do anything useful with the exported file without having to develop new software from scratch for the singular purpose of parsing it for hours, and am pretty angry about it. removed a bunch of counterproductive invective)

Hey :slight_smile: I got no problem with invectives :slight_smile:
The format of the export file is not user friendly, we know about that - it was mainly meant as a backup format back then and less for direct use. That’s why I was super happy when one of our users came up with an idea of a conversion tool into JSON which added a lot of usefulness, since we’re moved from an extremely nonstandard csv file with variable headers to a standard.
But I can see your point - JSON is friendly to programmers and Linux geeks, but probably not that much for casual spreadsheet visualisation.

But I will need your help here. You probably have some ideas in your head on what would be useful for you. Let’s start from there.
What do you think the format should look like? What should it show? What do you mean by “ready to use”?

Thanks for your grace on replying.
I think user-accessibility (as opposed to software-) may warrant a different standard than for pure data backup and restoration that relies on, e.g. records that always have the same number of fields in the same order (even if it’s inefficient because human/dumb-software parsing consistency is more important); one line of headers followed by records; and records that are generally limited to one line in length (insert rabbithole about multi-line fields).
I realise there are things that complicate the above, like if a user switches between accelerometer- and sonar-sensing, or turns snore recording on/off multiple times as I have: the records in that data don’t look like they all have the same structure beyond the initial common parts, they don’t all have the same length depending on what mode the user has the app in, and occasionally additional blank lines get thrown in, etc.
I imagine this also relates directly to every record having a new line of headers - each different recording type has its own format, so you list the format for every record to be sure?
Mine may be an isolated couple of usage-cases (testing both sonar and snore recording multiple times) but it’s an example of a structural difference you can spot just by trying to import the CSV directly into libreoffice calc, for example.

Thank you for your reply. I have not made my questions clear - I’d like to leave out the technical hurdles on the side for a while and focus on the end purpose.

What do you want to have in the end? A graph in Calc plotting the motion data? Or are you trying to plot trends in snoring? Or no graph, but rather some pivot table? Or other uses?

The data from Sleep as Android are multidimensional, which is where the real issue lies, you can’t fit all of them easily into a 2D table at the same time, so unconstrained formats like JSON are the only way to capture the whole.

My desired end would be a longitudinal graph of time (days) vs. hrs of the day to graph at minimum sleep+awake times to sketch long-term trends in diurnal rhythm. Ideally I’d like to also be able to use the accelerometer data to express activity over each strip of a day as bright vs. dark or contrasting colours in gradient with the ‘activity’ level recorded (the vertical axis of the graph in the app’s normal daily view if you’re using the accelerometer mode).
As soon as I started typing that I could see what you meant about multidimensional and realised that my expectations of what I could get out of a spreadsheet were half-baked. Spot the person who aborted a comp sci degree because they realised they’re mathematically-challenged. :B