A minimalistic GUI for analyzing Pandas DataFrames based on wxPython.
import dfgui
dfgui.show(df)- Tabular view of data frame
- Columns are sortable (by clicking column header)
- Columns can be enabled/disabled (left click on 'Columns' tab)
- Columns can be rearranged (right click drag on 'Columns' tab)
- Generic filtering: Write arbitrary Python expression to filter rows. Warning: Uses Python's
eval-- use with care. - Histogram plots
- Scatter plots
The default view: Nothing fancy, just scrolling and sorting. The value of cell can be copied to clipboard by right clicking on a cell.
The column selection view: Left clicking enables or disables a column in the data frame view. Columns can be dragged with a right click to rearrange them.
The filter view: Allows to write arbitrary Pandas selection expressions. The syntax is: An underscore _ will be replaced by the corresponding data frame column. That is, setting the combo box to a column named "A" and adding the condition _ == 1 would result in an expression like df[df["A"] == 1, :]. The following example filters the data frame to rows which have the value 669944 in column "UserID" and datetime.date value between 2016-01-01 and 2016-03-01.
Histogram view:
Scatter plot view:
Since wxPython is not pip-installable, dfgui does not handle dependencies automatically. You have to make sure the following packages are installed:
- pandas/numpy
- matplotlib
- wx
I haven't submitted dfgui to PyPI (yet), but you can install directly from git (having met all requirements). For instance:
git clone [email protected]:bluenote10/PandasDataFrameGUI.git dfgui
cd dfgui
pip install -e .
# and to check if everything works:
./demo.pyIn fact, dfgui only consists of a single module, so you might as well just download the file dfgui/dfgui.py.