Question rather than an issue

[joergheintz]

Hi Matt,

I wanted to use your package to investigate this data set here.

I installed Python 3.6.1 (3.6 produced errors), downgraded to scikit-learn==0.18.2 (20.xx produced errors), and could run the demo. It generated the hmtl file, and the feature csv file, Applying gaitpy on this dataset though doesn't work. The bouts can not be identified.

sample rate = 100Hz
size = 170cm
sample size = 670 records (below is a subset)

id         x          y         **z**      seq    **time**  gender   age

114431 0.040 -0.930 -0.068 1 10 1 25
114431 0.044 -0.922 -0.090 1 20 1 25
114431 0.052 -0.930 -0.100 1 30 1 25
114431 0.088 -0.954 -0.108 1 40 1 25
114431 0.108 -0.952 -0.118 1 50 1 25

So, I checked the demo materials you provided and couldn't find your input data. Could you provide the csv file with the actual accelerator data? I think this would help to narrow down the issue.

Command Line Output:
Classifying bouts of gait...
Bout classification complete!
Extracting features...
The classified_gait data indicates no bouts of gait were detected, aborting...
Plotting contacts...
Unable to load gait features: Please make sure the gait_features is in the correct format, aborting...

Best George

[matt002]

Hey George, Thanks for your email. The data used for the demo can be found in the demo_data.csv file in the GaitPy / gaitpy / demo folder on Github. https://github.com/matt002/GaitPy/blob/master/gaitpy/demo/demo_data.csv Based on the subset of data you show, I would make sure you specify the data is in units of g using the v_acc_units argument when creating an instance of GaitPy (v_acc_units='g'). Another possible source of error comes from your units of time - you may need to convert your units into Unix timestamps. If you do not know the actual time the data occurs - an arbitrary period of time can be specified. I hope that helps. Best of luck with your analysis! Best, Matt

…

On Tue, Jun 16, 2020 at 12:01 PM Joerg Heintz ***@***.***> wrote: Hi Matt, I wanted to use your package to investigate this data set here <http://www.am.sanken.osaka-u.ac.jp/BiometricDB/InertialGait.html>. I installed Python 3.6.1 (3.6 produced errors), downgraded to scikit-learn==0.18.2 (20.xx produced errors), and could *run the demo.* It generated the hmtl file, and the feature csv file, Applying gaitpy on this dataset though doesn't work. The *bouts can not be identified.* sample rate = 100Hz size = 170cm sample size = 670 records (below is a subset) id x y **z** seq **time** gender age 114431 0.040 -0.930 *-0.068* 1 *10* 1 25 114431 0.044 -0.922 *-0.090* 1 *20* 1 25 114431 0.052 -0.930 *-0.100* 1 *30* 1 25 114431 0.088 -0.954 *-0.108* 1 *40* 1 25 114431 0.108 -0.952 *-0.118* 1 *50* 1 25 So, I checked the demo materials you provided and couldn't find your input data. *Could you provide the csv file with the actual accelerator data?* I think this would help to narrow down the issue. *Command Line Output:* *Classifying bouts of gait... Bout classification complete! Extracting features... The classified_gait data indicates no bouts of gait were detected, aborting... Plotting contacts... Unable to load gait features: Please make sure the gait_features is in the correct format, aborting...* Best George — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#25>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEFCXQF4ZWABZD2RQQIPQKLRW6JNJANCNFSM4N7YHK7Q> .

[joergheintz]

Matt,

Thanks for the advise.

I worked with your module and applied it on the OSAKA data set (http://www.am.sanken.osaka-u.ac.jp/BiometricDB/InertialGait.html) 9m distance, 2 walks, 740 subjects from 3-78 years. and got some results you may be interested in. And I am curios if you find the results, findings plausible. https://rpubs.com/jheintz/651884. The entire document is under construction, please consider only tables and graphics. I am happy to schedule a zoom call if you want to discuss. George.