How to configure reports - grishasen/proof_of_value GitHub Wiki

Report Definitions

Purpose: Define how reports are displayed.

Each report includes:

  • metric: One of the configured metrics.
  • type: Visualization type.
  • description: User-facing text.
  • group_by: IH columns used for grouping.
  • Additional fields depend on report type (see below).

Report Types and Config Options:


Bar Polar Plot

Option Description
metric engagement, conversion
type bar_polar
group_by IH columns
r Metric value
theta Angular axis control
color Coloring metric
showlegend Show/hide legend

Gauge Plot

Option Description
metric engagement, conversion
type gauge
value Metric score
reference Baseline values per group

Treemap

Option Description
metric engagement, conversion, ml_scores
type treemap
group_by IH columns
color Color metric

Heatmap

Option Description
metric engagement, conversion, ml_scores
type heatmap
x, y Axis selections
color Metric for coloring

Line/Bar Plot

Option Description
metric Any category
type line, bar
x, y Axis fields
color Column for color differentiation
facet_row, facet_col Split subplots by grouping
metric_totals_toggle Show summary cards
advanced_options_toggle Enable extra config UI

Line Plot (Z-Score Variant)

| Special Logic | Z-scores in the -1.96..1.96 range highlighted (statistically insignificant, 95% CI). |


Line Plot (G-Test, Chi-Square)

Feature Interpretation
Red Control better
Yellow Insignificant
Green Test group better

Assumption: 2 groups only, first is Control.


Exploratory Data Analysis (EDA)

Purpose: Visualize IH dataset characteristics.

Metrics shown:

  • Count, Mean, Median, Percentiles (25, 75, 90)
  • Std, Skew for fields like Propensity, Priority

Customer Lifetime Value (CLV) Models

  • RFM Analysis:
    • Inputs: recency, frequency, monetary_value
    • Output: Segmentations like Top Spenders, At Risk, Premium