Closed alrod97 closed 4 years ago
is the dash plotting reimplementation ready?
is the dash plotting reimplementation ready? The new classes are currently being used in the dashboards.py file. Furthermore, my idea is to have an e.g a folder with different dashboards and then call/display a specific dashboard depending if the user wishes to use PCA, LLE, etc. I'm not quite sure how to do that, maybe someone has an idea. The current dashboard is for PCA.
As discussed, we'd like to split the big current EvaData class where all the PCA,TSNE is being calculated and also all the Plotly figures are being created into two independent classes.
The first class is DataStatistics(), you can find it in the statistics_methods.py file. It is a cleaner, more general version that is used for any math-statistical-ML computation. It is mainly though to work with pandas data frames. Therefore, if you wanna add any kind of new method (e.g KERNEL PCA), you just need to create a method inside the class and use the pandas data frame. Currently, we can use PCA, LLE or TSNE
The second class is VisualizationPlotly which is only meant to produce Plotly figures with an input pandas data frame. Depending on the user's wishes, we can plot the data as a 1D/2D/3D scatter plot, plot statistical information about the outliers/inliers in a Boxplot and plot an histogram for a given feature. Again, this class is meant to work with pandas data frames.