Add new charts

pulkka/chart_utils.py

@@ -1,9 +1,9 @@
 from bokeh import models as bm
 from bokeh import plotting as bp
 from bokeh.transform import factor_cmap
-from pandas import DataFrame
+from pandas import DataFrame, Series
 
-from pulkka.data_utils import get_categorical_stats
+from pulkka.data_utils import explode_multiselect, get_categorical_stats
 
 CAT_Q_RADIUS = 0.1
 
@@ -88,3 +88,27 @@ def get_categorical_stats_plot(df, *, category, value, na_as_category=None, line
             color="#B53471",
         )
     return plot
+
+
+def get_multiselect_frequency_plot(
+    series: Series,
+    *,
+    title: str,
+    top_n: int = 20,
+) -> bp.figure:
+    """Horizontal bar chart of the top N values from a comma-separated multiselect column."""
+    counts = explode_multiselect(series, top_n=top_n)
+    # Reverse so highest count is at the top
+    labels = list(counts.index[::-1])
+    values = list(counts.values[::-1])  # noqa: PD011
+
+    plot = bp.figure(
+        title=title,
+        y_range=labels,
+        height=max(300, 22 * len(labels)),
+        width=700,
+    )
+    plot.hbar(y=labels, right=values, height=0.7, color="#2a6180")
+    plot.xaxis.axis_label = "Vastauksia"
+    plot.x_range.start = 0
+    return plot

pulkka/generate_charts.py

@@ -7,6 +7,7 @@ from pulkka.chart_utils import (
     gender_colormap,
     get_categorical_stats_plot,
     get_df_hover_tool,
+    get_multiselect_frequency_plot,
     set_yaxis_cash,
 )
 from pulkka.column_maps import (
@@ -19,6 +20,17 @@ from pulkka.column_maps import (
 from pulkka.config import OUT_DIR
 from pulkka.data_ingest import read_data
 
+MULTISELECT_CHARTS = {
+    "Data & ML": "Data & ML (top 15)",
+    "DevOps & pilvi": "DevOps & pilvi (top 20)",
+    "Edut (ei luontoisedut)": "Edut (top 15)",
+    "Käyttöjärjestelmä": "Käyttöjärjestelmä",
+    "Luontoisedut": "Luontoisedut (top 15)",
+    "Ohjelmointikieli": "Ohjelmointikielet (top 20)",
+    "Tietokannat": "Tietokannat (top 15)",
+    "Web-kehykset": "Web-kehykset (top 20)",
+}
+
 plot_funcs = set()
 
 
@@ -76,9 +88,24 @@ def plot_kaupunki_vuositulot(df: DataFrame):
 def main():
     df = read_data()
     plots = [func(df) for func in sorted(plot_funcs, key=lambda f: f.__name__)]
+
     bp.output_file(OUT_DIR / "charts.html", title="Koodiklinikan Palkkakysely")
     bp.save(bl.grid(plots, ncols=2, sizing_mode="stretch_both"))
 
+    multiselect_plots = []
+    for col, title in MULTISELECT_CHARTS.items():
+        if col in df.columns:
+            top_n = 20 if "20" in title else 15
+            multiselect_plots.append(
+                get_multiselect_frequency_plot(df[col], title=title, top_n=top_n),
+            )
+
+    bp.output_file(
+        OUT_DIR / "charts2.html",
+        title="Koodiklinikan Palkkakysely – Monivalinnat",
+    )
+    bp.save(bl.grid(multiselect_plots, ncols=2, sizing_mode="stretch_both"))
+
 
 if __name__ == "__main__":
     main()