Data smoothing function

1,181 views
Skip to first unread message

nirakara

unread,
Nov 2, 2010, 11:59:47 AM11/2/10
to Sleep as an Droid
I like your app and I would love to use it more, but I'm particularly
interested in using the data and it's a bit hard to extract the data
right now (I have to copy it to an email, create my own x-values,
etc). But more qualitatively I'm interested in making the data look
like the graphs you show in the program. What kind of filtering/
convolution are you using to make it look so "nice"? Perhaps the
program could export the data filtered or as images? Thanks!

Petr Nalevka

unread,
Nov 2, 2010, 12:55:58 PM11/2/10
to sleep-as...@googlegroups.com, Martin Stava
Thank you very much for your feed-back on Sleep as an Droid. You can get sleep graph data in two ways. Clicking on the graph and choosing "Copy Data" gives you aggregated data as they are in the graph (only a corner path effect is applied to make it look more round, but this can Excel or Google Docs do also just by choosing the right chart). Second, there is a menu option "Copy latest raw data", and this gives raw data before aggregation and smoothening. 

You can also send the generated graph as images. But currently we only support FaceBook but we plan to allow sending the graphs using e-mail in the versions to come.  

What would be you preferred way of extracting the data?

Thank you.

Best regards,

Petr

Nicolás Sherwood Droz

unread,
Nov 2, 2010, 1:47:28 PM11/2/10
to sleep-as...@googlegroups.com, Martin Stava

Thanks for the quick reply Petr! I think you have the best android alternative to the apple sleep cycle app right now.

However from what you said maybe there is a bug because I get literally the same data if I "copy data" from the specific graph or if I "copy latest raw data" from the parent menu. 

Aside from that it would be cool if you could export the data directly into a csv file, or since you mentioned google docs maybe an option that directly puts it into a google spreadsheet, so you don't have to use another program to move the data around. That would be pretty sweet. And if we're comparing with the other program two cute features you might consider are 1) A "Test button" which makes a sound when you shake the phone and you can show your friends how it works, and 2) Labels or ticks on the graph image that show where the deep sleep, dreaming, and awake points are (even if they're not completely accurate). But those last two are just suggestions to make your app more popular among the masses :)

Also can you expand on the "corner path effect" you mentioned? Is there a method for that in the android library or did you use a math function to do it? I just can't get my graphs to be as nice as yours no matter what moving average window I use.

Thanks so much!

Nick

Petr Nalevka

unread,
Nov 3, 2010, 11:06:56 AM11/3/10
to sleep-as...@googlegroups.com, Martin Stava
Hello Nicolas,

please see my replies below.

2010/11/2 Nicolás Sherwood Droz <nic...@gmail.com>


However from what you said maybe there is a bug because I get literally the same data if I "copy data" from the specific graph or if I "copy latest raw data" from the parent menu. 
This may happen when your sleep graph is shorter than 100 entries. In this case raw data are identical with the stared graph data. when longer aggregation takes place and those two start to differ.

Aside from that it would be cool if you could export the data directly into a csv file, or since you mentioned google docs maybe an option that directly puts it into a google spreadsheet, so you don't have to use another program to move the data around. That would be pretty sweet.
We are definitely considering this feature.
 
And if we're comparing with the other program two cute features you might consider are 1) A "Test button" which makes a sound when you shake the phone and you can show your friends how it works, and 2) Labels or ticks on the graph image that show where the deep sleep, dreaming, and awake points are (even if they're not completely accurate). But those last two are just suggestions to make your app more popular among the masses :)

Those features are definitely cool, thank you for your suggestions. We are currently working on the labels, but there are issue as the data being measured are relative. 

Also can you expand on the "corner path effect" you mentioned? Is there a method for that in the android library or did you use a math function to do it? I just can't get my graphs to be as nice as yours no matter what moving average window I use.

Reply all
Reply to author
Forward
0 new messages