Jupyter Notebook to graph summarised data 4 dimensionally.
Inspired by the Bokeh Crossfilter Example.
The template works best with the anaconda python distribution.
To copy the exact environment used to create this, create a conda environment from the yml file in this repo.
Click here for a how-to.
-
Once opened, enter the location of the the csv you'd like to visualise.
-
Select the discrete (non-numeric) variables.
-
Select preferred settings, e.g. variables, colour scheme, circle size etc.
-
To save a graph as a static png, click the Save button in the top-right of the figure. To select an individual data point, click the linked table or hover over the circle to see tooltip information.
- Ensure each unique variable in the data is a column.
- This is especially important for time-series data.
- Add a Binder for this
- Add log scale option
- File import wizard
- Not currently possible.
- Add 'none' as an option for default variables
- Refactor code to avoid the use of global scope, return tuple instead (good practice).
- Change to 1-based indexing for row number tooltip?
- Move as many widgets to Bokeh from iPython without making it clunky
- Add all columns to data table
- Change continuous to categorical data,
- Dropdown linked with slider? This may make it too complicated.