correlation start

02_quant_analysis.py

@@ -14,25 +14,27 @@ def _():
     from utils import JPMCSurvey, combine_exclusive_columns, calculate_weighted_ranking_scores
     from plots import plot_average_scores_with_counts, plot_top3_ranking_distribution, plot_ranking_distribution, plot_most_ranked_1, plot_weighted_ranking_score, plot_voice_selection_counts, plot_top3_selection_counts
 
-    import plots as plts
-    import utils as utl
+    import plots
+    import utils
+
+    from speaking_styles import SPEAKING_STYLES
     return (
         JPMCSurvey,
         Path,
+        SPEAKING_STYLES,
         calculate_weighted_ranking_scores,
         check_progress,
         duration_validation,
         mo,
         pl,
-        plot_average_scores_with_counts,
         plot_most_ranked_1,
         plot_ranking_distribution,
         plot_top3_ranking_distribution,
         plot_top3_selection_counts,
         plot_voice_selection_counts,
         plot_weighted_ranking_score,
-        plts,
-        utl,
+        plots,
+        utils,
     )
 
 
@@ -47,7 +49,7 @@ def _():
 def _(JPMCSurvey, QSF_FILE, RESULTS_FILE):
     survey = JPMCSurvey(RESULTS_FILE, QSF_FILE)
     data_all = survey.load_data()
-    data_all.collect()
+    # data_all.collect()
     return data_all, survey
 
 
@@ -298,7 +300,7 @@ def _(mo):
 
 
 @app.cell
-def _(data, survey, utl):
+def _(data, survey, utils):
     ss_or, choice_map_or = survey.get_ss_orange_red(data)
     ss_gb, choice_map_gb = survey.get_ss_green_blue(data)
 
@@ -309,12 +311,12 @@ def _(data, survey, utl):
     choice_map = {**choice_map_or, **choice_map_gb}
     # print(_d.head())
     # print(choice_map)
-    ss_long = utl.process_speaking_style_data(ss_all, choice_map)
-    return (ss_long,)
+    ss_long = utils.process_speaking_style_data(ss_all, choice_map)
+    return choice_map, ss_all, ss_long
 
 
 @app.cell
-def _(mo, pl, plts, ss_long):
+def _(mo, pl, plots, ss_long):
     content = """### How does each voice score for each “speaking style labeled trait”?"""
 
     for i, trait in enumerate(ss_long.select("Description").unique().to_series().to_list()):
@@ -323,7 +325,7 @@ def _(mo, pl, plts, ss_long):
         content += f"""
     ### {i+1}) {trait.replace(":", " ↔ ")}
 
-    {mo.ui.plotly(plts.plot_speaking_style_trait_scores(trait_d, title=trait.replace(":", " ↔ "), height=550))}
+    {mo.ui.plotly(plots.plot_speaking_style_trait_scores(trait_d, title=trait.replace(":", " ↔ "), height=550))}
     """
 
     mo.md(content)
@@ -339,17 +341,17 @@ def _(mo):
 
 
 @app.cell
-def _(data, mo, plot_average_scores_with_counts, survey):
+def _(data, mo, plots, survey):
     vscales = survey.get_voice_scale_1_10(data)[0].collect()
-    plot_average_scores_with_counts(vscales, x_label='Voice', width=1000)
+    # plot_average_scores_with_counts(vscales, x_label='Voice', width=1000)
 
     mo.md(f"""
 
     ### How does each voice score on a scale from 1-10?
 
-    {mo.ui.plotly(plot_average_scores_with_counts(vscales, x_label='Voice', width=1000))}
+    {mo.ui.plotly(plots.plot_average_scores_with_counts(vscales, x_label='Voice', width=1000))}
     """)
-    return
+    return (vscales,)
 
 
 @app.cell(hide_code=True)
@@ -373,16 +375,57 @@ def _(mo):
     return
 
 
-@app.cell
+@app.cell(hide_code=True)
 def _(mo):
     mo.md(r"""
-    ### Total Results
+    ### How to Interpret These Correlation Results
+    Each bar represents the Pearson correlation coefficient (r) between a speaking style trait rating (1-5 scale) and the overall Voice Scale rating (1-10).
 
-    - [ ] 4 correlation diagrams
+    **Reading the Chart**
+
+    | Correlation Value |	Interpretation |
+    |-----------|----------|
+    | r > 0 (Green bars)| 	Positive correlation — voices rated higher on this trait tend to receive higher Voice Scale scores|
+    | r < 0 (Red bars)| 	Negative correlation — voices rated higher on this trait tend to receive lower Voice Scale scores|
+    | r ≈ 0| 	No relationship — this trait doesn't predict Voice Scale ratings|
     """)
     return
 
 
+@app.cell
+def _(choice_map, ss_all, utils, vscales):
+    df_style = utils.process_speaking_style_data(ss_all.collect(), choice_map)
+    df_voice_long = utils.process_voice_scale_data(vscales)
+
+    joined_df = df_style.join(df_voice_long, on=["_recordId", "Voice"], how="inner")
+    # df_voice_long
+    return df_style, joined_df
+
+
+@app.cell
+def _(SPEAKING_STYLES, joined_df, mo, plots):
+    _content = """### Total Results
+
+    """
+
+    for style, traits in SPEAKING_STYLES.items():
+        # print(f"Correlation plot for {style}...")
+        fig = plots.plot_speaking_style_correlation(
+            df=joined_df,
+            style_color=style,
+            style_traits=traits,
+            title=f"Correlation: Speaking Style {style} and Voice Scale 1-10"
+        )
+        _content += f"""
+    #### Speaking Style **{style}**:
+    
+    {mo.ui.plotly(fig)}
+
+    """
+    mo.md(_content)
+    return
+
+
 @app.cell
 def _(mo):
     mo.md(r"""
@@ -425,6 +468,30 @@ def _(mo):
     return
 
 
+@app.cell
+def _(SPEAKING_STYLES, df_style, mo, plots, top3_voices, utils):
+    df_ranking = utils.process_voice_ranking_data(top3_voices)
+    joined = df_style.join(df_ranking, on=['_recordId', 'Voice'], how='inner')
+
+
+    _content = """## Correlations Voice Speaking Styles <-> Voice Ranking Points
+
+    """
+
+    for _style, _traits in SPEAKING_STYLES.items():
+        _fig = plots.plot_speaking_style_ranking_correlation(joined, _style, _traits)
+        _content += f"""
+
+        #### Speaking Style **{_style}**:
+    
+        {mo.ui.plotly(_fig)}
+    
+        """
+
+    mo.md(_content)
+    return
+
+
 @app.cell(hide_code=True)
 def _(mo):
     mo.md(r"""