Plot Input (`sketch.plot_input`)

This module provides input-data visualisation helpers for metrics, correlation structure, and target-over-time views.

Import style:

from sketch.plot_input import plot_all_metrics, plot_correlation_matrix, plot_all_media_spend
# Compatibility form in some environments:
# from src.sketch.plot_input import ...

plot_all_metrics(input_data, output_dir, suffix)

Plots all media volumes/costs, extra features, and target in a multi-panel Matplotlib figure.

Saved artefact:

• metrics_{suffix}.png → 10_pre_diagnostics/

The function persists via save_figure(..., stage=STAGE_PRE_DIAGNOSTICS).

plot_correlation_matrix(input_data, per_observation_df)

Returns a Plotly heatmap and correlation DataFrame for volume columns plus target.

• Returns: (fig, corr_df)
• No automatic file save.

If you want persistence, save manually into your chosen stage directory (for example 10_pre_diagnostics/) with Plotly I/O methods.

plot_all_media_spend(input_data, per_observation_df)

Returns a Plotly line chart of target over time.

• Returns: fig
• No automatic file save.

As above, save manually if needed.

Usage example

from sketch.plot_input import plot_all_metrics, plot_correlation_matrix

plot_all_metrics(input_data, output_dir=results_dir, suffix="raw")

corr_fig, corr_df = plot_correlation_matrix(input_data, per_observation_df)
# Optional manual save:
# corr_fig.write_html(f"{results_dir}/10_pre_diagnostics/correlation_matrix.html")