added top_n_frequency_analysis, and rotating ticks on barplot, countplot

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "studioai"
-version = "0.2.21"
+version = "0.2.22"
 description = "Atelier for Artificial Intelligence and Data Science"
 authors = [
     "John James <[email protected]>",

setup.cfg

@@ -33,6 +33,6 @@ per-file-ignores =
 
 # Enables maccabe complexity checks
 #   see https://github.com/PyCQA/mccabe#plugin-for-flake8
-max-complexity = 10
+max-complexity = 15
 
 exclude = .git,__pycache__,old,build,dist,.venv,.eggs,.tox

studioai/analysis/explore/eda.py

@@ -11,7 +11,7 @@
 # URL        : https://github.com/john-james-ai/studioai                                           #
 # ------------------------------------------------------------------------------------------------ #
 # Created    : Thursday August 10th 2023 08:29:08 pm                                               #
-# Modified   : Wednesday May 22nd 2024 02:54:40 am                                                 #
+# Modified   : Friday June 7th 2024 02:16:21 pm                                                    #
 # ------------------------------------------------------------------------------------------------ #
 # License    : MIT License                                                                         #
 # Copyright  : (c) 2023 John James                                                                 #
@@ -65,7 +65,7 @@ def plot(self) -> Visualizer:  # pragma: no cover
         return self._visualizer
 
     @property
-    def stats(self) -> Visualizer:  # pragma: no cover
+    def stats(self) -> Inference:  # pragma: no cover
         self._inference.data = self.df
         return self._inference
 
@@ -256,6 +256,66 @@ def unique(self, columns: list = None) -> pd.DataFrame:
             df = self.df.drop_duplicates().reset_index(drop=True)
         return self._format(df=df)
 
+    # ------------------------------------------------------------------------------------------- #
+    def top_n_frequency_analysis(self, x: str, n: int, data: pd.DataFrame = None):
+        """Returns a dataframe with proportional and cumulative counts of one or more categorical variables.
+
+        Args:
+            x (Union[str,List[str]]): A string or list of strings indicating the variables included in the count.
+            n (int): Number of rows to include in top n.
+            data (pd.DataFrame). Data to analyze. Optional.
+
+        """
+
+        # Use instance variable df if data is None
+        data = data or self.df
+
+        # Calculate frequency distribution
+        freq = data[x].value_counts().reset_index()
+        freq.columns = [x, "Count"]
+
+        # Calculate cumulative count and proportions
+        freq["Cumulative Count"] = freq["Count"].cumsum()
+        total_count = freq["Count"].sum()
+        freq["Proportion"] = freq["Count"] / total_count
+        freq["Cumulative Proportion"] = freq["Proportion"].cumsum()
+
+        # Top N rows
+        top_n = freq.head(n).copy()
+
+        # Row for the rest of the dataset
+        if len(freq) > n:
+            rest_count = freq.iloc[n:]["Count"].sum()
+            rest_cumulative_count = top_n["Cumulative Count"].iloc[-1] + rest_count
+            rest_proportion = rest_count / total_count
+            rest_cumulative_proportion = (
+                1.0  # because it's the rest, it covers the remaining percentage
+            )
+            rest_row = pd.DataFrame(
+                {
+                    x: [f"Rest of {x}"],
+                    "Count": [rest_count],
+                    "Cumulative Count": [rest_cumulative_count],
+                    "Proportion": [rest_proportion],
+                    "Cumulative Proportion": [rest_cumulative_proportion],
+                }
+            )
+            top_n = pd.concat([top_n, rest_row], ignore_index=True)
+
+        # Total row
+        total_row = pd.DataFrame(
+            {
+                x: ["Total"],
+                "Count": [total_count],
+                "Cumulative Count": [total_count],
+                "Proportion": [1.0],
+                "Cumulative Proportion": [1.0],
+            }
+        )
+        top_n = pd.concat([top_n, total_row], ignore_index=True)
+
+        return top_n
+
     # ------------------------------------------------------------------------------------------- #
     def frequency(
         self,

studioai/analysis/visualize/visualizer.py

@@ -11,7 +11,7 @@
 # URL        : https://github.com/john-james-ai/studioai                                           #
 # ------------------------------------------------------------------------------------------------ #
 # Created    : Saturday August 26th 2023 06:25:27 am                                               #
-# Modified   : Thursday May 9th 2024 09:26:13 am                                                   #
+# Modified   : Friday June 7th 2024 02:18:54 pm                                                    #
 # ------------------------------------------------------------------------------------------------ #
 # License    : MIT License                                                                         #
 # Copyright  : (c) 2023 John James                                                                 #
@@ -381,6 +381,8 @@ def countplot(
         plot_counts: bool = False,
         title: str = None,
         figsize: bool = (12, 4),
+        rotate_xticks: int = None,
+        rotate_yticks: int = None,
         ax: plt.Axes = None,
         **kwargs,
     ) -> plt.Axes:
@@ -402,6 +404,8 @@ def countplot(
             plot_counts (bool): If True, the bars are annotated with absolute and relative counts. Default = False
             title (str): Title for the plot. Optional
             figsize (tuple): Size of figure in inches. Ignored if ax is provided.
+            rotate_xticks (int): The degrees to rotate the x_ticks. None means no rotation.
+            rotate_yticks (int): The degrees to rotate the y_ticks. None means no rotation.
             ax: (plt.Axes): A matplotlib Axes object. Optional. If not provide, one will be obtained from the canvas.
 
         """
@@ -436,6 +440,22 @@ def countplot(
             *args,
             **kwargs,
         )
+
+        if rotate_xticks is not None:
+            ax = ax.set_xticks(
+                ax.get_xticks(),
+                ax.get_xticklabels(),
+                rotation=rotate_xticks,
+                ha="right",
+            )
+        elif rotate_yticks is not None:
+            ax = ax.set_yticks(
+                ax.get_yticks(),
+                ax.get_yticklabels(),
+                rotation=rotate_yticks,
+                va="center",
+            )
+
         if plot_counts:
             if orient == "v":
                 for p in ax.patches:
@@ -472,6 +492,8 @@ def barplot(
         orient: str = None,
         title: str = None,
         figsize: bool = (12, 4),
+        rotate_xticks: int = None,
+        rotate_yticks: int = None,
         ax: plt.Axes = None,
         **kwargs,
     ) -> plt.Axes:
@@ -493,6 +515,8 @@ def barplot(
                 when both x and y are numeric or when plotting wide-form data.
             title (str): Title for the plot. Optional
             figsize (tuple): Size of figure in inches. Ignored if ax is provided.
+            rotate_xticks (int): The degrees to rotate the x_ticks. None means no rotation.
+            rotate_yticks (int): The degrees to rotate the y_ticks. None means no rotation.
             ax: (plt.Axes): A matplotlib Axes object. Optional. If not provide, one will be obtained from the canvas.
 
         """
@@ -515,6 +539,21 @@ def barplot(
             **kwargs,
         )
 
+        if rotate_xticks is not None:
+            ax = ax.set_xticks(
+                ax.get_xticks(),
+                ax.get_xticklabels(),
+                rotation=rotate_xticks,
+                ha="right",
+            )
+        elif rotate_yticks is not None:
+            ax = ax.set_yticks(
+                ax.get_yticks(),
+                ax.get_yticklabels(),
+                rotation=rotate_yticks,
+                va="center",
+            )
+
         if title is not None:
             _ = ax.set_title(title)
 
@@ -1286,8 +1325,8 @@ def heatmap(
         cbar_kws: dict = None,
         cbar_ax: plt.Axes = None,
         square: bool = True,
-        xticklabels: Union["str, bool", list, int] = "auto",
-        yticklabels: Union["str, bool", list, int] = "auto",
+        xticklabels: Union["str", "bool", list, int] = "auto",
+        yticklabels: Union["str", "bool", list, int] = "auto",
         mask: bool = None,
         ax: plt.Axes = None,
         title: str = None,