ssmode 0.1.6

Tools for use in Mode Analytics Jupyter Notebooks

ssmode

Installing & Importing

To use the helper functions in a Mode Python Notebook, first install the package by adding this cell:

pip install ssmode -t "/tmp" > /dev/null 2>&1

You can then import various functions like:

from ssmode.tables import style_table

Styling Documentation

The functions in files bar_chart.py, kpi.py, tables.py are designed to "fake" the Mode Analytics built-in widgets from the Notebook.

Styling Tables

style_table(df, hl_type=None, n=3, bar_cols=[])

• df: Pandas Dataframe (index will not be displayed)
• hl_type: None or string 'nlargest', 'gradient' or 'bars' specifying cell highlighting type
• n: integer > 0 specifying how many greatest cells will be highlighted (only applicable for hl_type='nlargest'), all numeric columns will get this style
• bar_cols: array of strings specifying which columns will get the "bar charty style" (only applicable for hl_type='bars')

Styling Bar & Line Charts

style_bar_chart(ptl_fig, ytitle='') OR style_line_chart(ptl_fig, ytitle='')

• ptl_fig: plotly chart with bars
• ytitle: title on y-axis

Displaying KPI Widget

display_as_kpi(kpi_name, value)

• kpi_name: string specifying KPI title/name (displayed on top)
• value: value of the KPI (large value)

Processing Functions Documentation

Outlier Removal

prune_quotes(df, variable_col, group_cols, log_scale=True, k=1.5, max_diffs=[(2,3),(3,5),(4,10)])

• df: Pandas Dataframe with data to remove outliers from
• variable_col: string with column name based on which outliers will be removed (e.g. 'price')
• group_cols: array of string(s) with columns to group df by for quartile calculation purposes (e.g. ['item_id'])
• log_scale: boolean specifying whether to use logarithmic scale for outlier removal
• k: float specifying limit ranges Q1-k*IQR and Q3+k*IQR
• max_diffs: list of tuples with two values, each tuple specifies total number of quotes and the maximal allowed ratio between max/min quotes (if violated, RFQ will be removed before the IQR outlier detection method)