NeuroDataDesign / orange-panda-f16s17

Automated EEG data analysis.
5 stars 5 forks source link

For May 1 #62

Closed rmarren1 closed 7 years ago

rmarren1 commented 7 years ago

Monday Commitments

Status Task Deliverable
:panda_face: Re-do pipeline schematic to fit on a single page and look not bad. lucid chart @rmarren1
Finish group analysis code and show compelling results justifying our changes to the pipeline. ssi exploration @rmarren1
:panda_face: Fix cloudy code, make more cost efficient. code here, success pictures @nkumarcc
Document cloud and any code possible. @nkumarcc
:panda_face: Figures to track computational resources used by pipeline, choose efficient EC2 instance size according to this. notebook with sample dataset, EC2 Instance Sizes @nkumarcc

Group

Bandstop is bad

After completion of the group analysis code, we noticed that bandstop filtering appears to be lowering discriminibility of the data significantly. (Here, you can ignore the non-blue lines)

old_group

No Bandstop with Spherical Splines Interpolation

ssi

No Bandstop, Wavelet Interpolation

wave

No Bandstop,

group_analysis

No Interpolation