Update dataset preview page
Update the dataset preview page with more data analysis plots.
(Notes from our initial pass at this are in issue #209, in particular some ui ideas are here https://github.com/EpistasisLab/pennai/issues/209#issuecomment-503251165 )
All changes can be seen on the data-preview branch.
A list of the new components / changes to existing components is as follows:
-
DatasetTabbed : Contains 3 tabs.
- The Preview tab contains the original data preview content (Dataset component)
- The Basic Stats tab contains newly added Plotly graphs (DataStats component)
- The Sample tab contains sample plots using custom test data (code can be seen in the parent component i.e. DatasetTabbed component)
-
DataStats component :
- Modifies data from row-wise format received from Papa parse to column-wise format (currently top 1000 rows are only selected for the graphs)
- Creates a list of categorical, ordinal and numerical features (slices to select first 20 from each category)
- Currently it only makes use of PlotlyBarPlot and PlotlyBoxPlot components.
- For future scope, we can separate data into multiple parts for each class and make use of PlotlyStackedBarPlot
-
PlotlyBarPlot : uses histogram count method to generate histograms, can be updated to any other custom function (see example usage in sample tab)
-
PlotlyBoxPlot : customized to stack one or more box plots (see example usage in sample tab)
-
PlotlyStackedBarPlot : stacked histogram plot (see example usage in sample tab)