danieldondorp / behaviouranalysis Goto Github PK

The common dataframe is now dumped as dataframes/common_data.pickle in the project folder, while it should be in the dataframes folder. This is a path formatting issue in the linux branch.

I suggest reorganizing stimuli-related columns for easy grouping and comparisons across different stimuli groups:
-Current ‘stimuli’ column should be renamed to ‘number_of_stimuli’
-New column called ‘stimuli’ or ‘stimuli class’ or similar should be made, that reads ‘none’ for the data without stimulation, and for data with light stimuli the class should be extracted and labeled as ‘green’, ‘red’, ‘blue’, ‘white’ for simple single-stimuli-single-color experiments. Label ‘mix’/’complex’ or similar for complex and/or wrong stimuli (i.e. more than 1 stimuli and/or color) for now. In the future you can also create new vibration classes, combined classes etc.
-If future experiments have stimuli of differing temporal composition, this can be added as a ‘stimuli_sequence’ column to sort your 1min stimuli from your 5-times-5seconds stimuli of same colour/vibration intensity etc.
-The deltas can then be reduced to just ‘delta_(ON/OFF)_(name of column)’ and compared between different stimuli settings (+ in the case of several stimuli in same experiment maybe add counter to deltas?)

For parameters with values centred around 0 (e.g. accelerations) a variance metaparameter would be very useful (either variance or standard deviation), since that is more likely to vary between groups than range and mean values.
We also need additional parameters for turn values. Calculate mean_ and var_turns as now, but only counting in values at swimming speeds (>500).

It makes more sense to have the code to generate plots in the PlotWidget class, and rather slice out and send over data from the PandasViewer class.