Add relative package ownership plot

notebooks/popcorn.py

@@ -158,8 +158,10 @@ def _():
 
 @app.cell
 def _(df_pkg_lazy: pl.LazyFrame):
+    pkg_per_day = df_pkg_lazy.group_by("date").agg(pl.col("count").sum()).sort("date")
+
     def _():
-        weekly_packages = df_pkg_lazy.group_by_dynamic("date", every="1w").agg(
+        weekly_packages = pkg_per_day.group_by_dynamic("date", every="1w").agg(
             pl.col("count").sum()
         )
         return (
@@ -179,6 +181,34 @@ def _(df_pkg_lazy: pl.LazyFrame):
     return
 
 
+@app.cell
+def plt_pkg_relative(pkg_per_day: pl.LazyFrame, df_unique_installs: pl.DataFrame):
+    def _():
+        relative_packages = (
+            pkg_per_day.with_columns(df_unique_installs["unique"])
+            .with_columns((pl.col("count") / pl.col("unique")).alias("relative"))
+            .group_by_dynamic("date", every="1w")
+            .agg(pl.col("relative").mean())
+        )
+        return (
+            lp.ggplot(
+                relative_packages.collect(engine="streaming"),
+                lp.aes("date", "relative"),
+            )
+            + lp.geom_line()
+            + lp.geom_smooth(method="loess")
+            + lp.labs(
+                title="Package ownership per user",
+                subtitle="Average relative weekly package ownership",
+                caption="Calculated by total amount of packages per day over unique installations",
+                y="number of packages",
+            )
+        )
+
+    _()
+    return
+
+
 @app.cell(hide_code=True)
 def _():
     mo.md(
@@ -256,12 +286,13 @@ def _(df_pkg_lazy: pl.LazyFrame):
 
 @app.cell
 def _():
-    (
+    df_unique_installs = pl.read_csv(
-        lp.ggplot(
-            pl.read_csv(
         f"{DATA_DIR}/unique_installs.csv",
         schema={"date": pl.Date, "unique": pl.UInt16},
-            ),
+    )
+    (
+        lp.ggplot(
+            df_unique_installs,
             lp.aes("date", "unique"),
         )
         + lp.geom_line()