⭐ All relevant columns should be logged and the main df should be stripped to these

ccp-eva / eyewit

👁‍🗨 Bootstrapping common eye-tracking tasks

MIT License

7 stars 3 forks source link

⭐ All relevant columns should be logged and the main df should be stripped to these #17

Closed kalaschnik closed 4 years ago

kalaschnik commented 4 years ago

Meassure execution time if df is large vs small

seems like microbenchmark() is measure to go

Mandatory columns for version 1

(Trial) is generated

Need to be in the raw data:

StudioEventData
GazePointX..ADCSpx. / check the other
GazePointY..ADCSpx. / check the other
RecordingTimestamp
FixationIndex
GazeEventDuration
StudioEventData
StudioEvent

kalaschnik commented 4 years ago

There is really no performance gain when stripping down the raw dataframe into a smaller datafame (column-wise). This is also true when creating new dataframes based on existing ones, where new dfs have new column... weird but good.

for more info use this commit for comparison tests: 0f4e9bc69105369bbd586e4420495e6551ed8e5e

run with microbenchmark(source("script.r"), source(script_full.R), times = 10)