py-geodetector 0.1.1__py3-none-any.whl

py_geodetector/__init__.py

@@ -0,0 +1 @@
+from .geodetector import GeoDetector, load_example_data

py_geodetector/example_data/disease.csv

@@ -0,0 +1,186 @@
+incidence,type,region,level
+5.94,7,5,5
+5.87,5,5,5
+5.92,5,5,5
+6.32,1,7,1
+6.49,3,2,4
+6.46,3,2,4
+6.51,3,2,4
+6.7,3,2,4
+6.68,3,2,4
+6.65,3,2,4
+6.65,3,2,4
+6.6,3,2,4
+6.66,3,2,4
+5.86,5,5,5
+6.93,3,2,4
+6.5,3,2,4
+6.37,3,2,4
+6.64,3,2,4
+6.43,7,9,1
+7.01,2,9,1
+6.77,2,9,1
+6.36,7,9,1
+6.44,7,9,3
+6.52,7,9,1
+5.83,5,6,7
+6.32,7,9,1
+6.33,7,9,1
+6.46,7,9,1
+6.63,7,9,6
+6.53,3,9,1
+6.59,7,9,4
+6.58,2,9,1
+6.63,3,9,6
+6.61,2,9,1
+7.04,3,9,4
+5.89,5,6,5
+6.57,2,9,1
+6.53,3,9,4
+6.58,3,9,4
+6.53,3,9,1
+6.55,3,9,4
+6.57,2,9,1
+6.57,1,7,1
+6.49,2,7,4
+6.56,2,9,4
+6.57,1,7,1
+5.97,5,6,5
+6.54,2,7,4
+6.71,3,9,6
+6.58,3,9,1
+6.63,7,9,6
+7.42,3,2,6
+6.71,3,4,1
+6.82,3,2,6
+6.56,3,2,6
+6.82,3,2,6
+5.86,5,6,7
+6.8,3,2,6
+7.74,3,2,6
+6.92,3,2,6
+6.75,3,2,6
+7.06,3,2,6
+7.1,3,2,6
+6.69,3,4,6
+6.81,3,2,6
+6.99,3,2,6
+6.79,3,2,6
+5.98,5,6,5
+7.01,3,2,6
+7.2,3,2,6
+6.95,3,2,4
+6.74,3,4,6
+6.77,3,4,1
+6.73,3,4,6
+6.57,7,3,3
+7.55,7,3,3
+6.04,1,6,5
+6.28,7,3,3
+6.38,7,3,3
+6.21,7,3,3
+6.33,7,3,3
+6.19,7,3,3
+6.45,7,3,2
+6.19,7,3,2
+6.14,7,3,2
+6.2,7,3,2
+6.14,3,3,2
+5.95,1,5,5
+6.27,7,3,3
+6.14,3,3,2
+6.18,7,1,2
+6.09,7,1,2
+6.04,7,1,3
+6.01,7,5,2
+6.01,7,5,3
+6.09,7,1,3
+5.85,5,6,5
+5.88,7,5,5
+5.8,5,6,7
+5.78,5,6,7
+5.7,5,6,7
+5.82,5,6,7
+5.73,5,6,7
+5.79,5,6,7
+5.8,5,6,7
+6.45,1,8,3
+6.47,1,8,3
+6.11,3,8,1
+5.98,7,5,5
+6.12,1,8,3
+6.15,1,8,1
+6.04,3,8,1
+6.06,3,8,3
+5.95,3,8,1
+6.09,3,8,1
+6.11,3,8,1
+6.27,1,8,1
+6.28,1,8,1
+6.28,1,8,1
+5.96,1,5,5
+6.23,1,8,1
+6.53,1,8,1
+6.25,1,8,3
+6.14,3,1,3
+6.48,3,1,3
+6.02,1,1,3
+6.03,1,1,3
+6.13,3,1,3
+6.07,3,1,3
+6.09,3,1,3
+5.66,5,5,5
+6.2,3,1,1
+6.13,1,1,3
+5.99,1,1,3
+6.13,1,1,1
+6.01,7,1,3
+6.11,1,1,3
+6.17,1,1,1
+6.19,3,1,1
+6.18,1,1,3
+6.09,1,1,3
+5.74,5,5,5
+6.19,7,1,3
+6.29,1,1,3
+6.25,1,1,1
+6.24,1,1,1
+6.33,3,1,1
+6.23,3,1,3
+6.35,1,8,1
+6.32,1,1,1
+6.37,1,1,1
+6.21,1,1,1
+5.88,5,5,5
+6.2,1,1,1
+6.58,1,8,1
+6.44,1,8,1
+6.48,1,8,1
+7.4,2,8,1
+6.38,1,7,1
+6.44,1,8,1
+6.65,1,8,1
+6.86,1,8,1
+6.71,1,3,1
+6.1,5,5,5
+6.88,7,3,1
+6.72,7,3,1
+6.3,7,3,1
+6.65,1,7,1
+6.34,1,8,1
+7.22,1,3,1
+6.5,1,3,3
+6.52,1,3,1
+6.53,7,3,1
+6.43,7,3,1
+5.89,5,5,5
+6.5,7,3,3
+6.48,7,3,3
+6.5,7,3,1
+6.38,1,7,1
+6.18,3,1,3
+6.39,1,7,1
+6.45,1,8,1
+6.86,2,7,1
+6.29,2,7,1
+6.49,1,7,1

py_geodetector/geodetector.py

@@ -0,0 +1,191 @@
+import warnings
+import numpy as np
+import pandas as pd
+from typing import Sequence
+from scipy.stats import f, ncf
+import matplotlib.pyplot as plt
+
+
+from pathlib import Path
+def load_example_data():
+    file_path = Path(__file__).parent / "example_data" / "disease.csv"
+    df = pd.read_csv(file_path)
+    return df
+
+
+def _plot_value(ax, interaction_df, ecological_df, value_fontsize=10):
+    length = len(interaction_df.index)
+    for i in range(length):
+        for j in range(length):
+            if not pd.isna(interaction_df.iloc[i, j]):
+                num = str(round(interaction_df.iloc[i, j], 2))
+                mark = num[-2:] if 3 == len(num) else num[-3:]
+                if 'Y'==ecological_df.iloc[i, j]:
+                    ax.text(j, i, mark, ha="center", va="center", color="r", fontsize=value_fontsize)
+                else:
+                    ax.text(j, i, mark, ha="center", va="center", color="k", fontsize=value_fontsize)
+
+
+class GeoDetector(object):
+    def __init__(self, df: pd.DataFrame, y: str, factors: Sequence[str]):
+        self.df = df
+        self.y = y
+        self.factors = factors
+        self._check_data(df, y, factors)
+        self.factor_df, self.interaction_df, self.ecological_df = None, None, None
+
+    def _check_data(self, df, y, factors):
+        for factor in factors:
+            if not factor in df.columns:
+                raise ValueError('Factor [{}] is not in data')
+            
+        for factor in factors:
+            # 检查列的数据类型
+            if df[factor].dtype not in ['int64', 'int32', 'int16', 'int8', 
+                                        'uint64', 'uint32', 'uint16', 'uint8', 
+                                        'object', 'string']:
+                # 如果数据类型不是整型或字符型，发出警告
+                warnings.warn(f"Factor '{factor}' is not of type 'int' or 'str'.")
+        
+        if y not in df.columns:
+            raise ValueError('Factor [{}] is not in data')
+            
+        for factor in factors:
+            if y==factor:
+                raise ValueError("Y variable should not in Factor variables. ")
+        
+        has_null = df.isnull().values.any()
+        if has_null:
+            raise ValueError("data hava some objects with value NULL")
+
+    @classmethod
+    def _cal_ssw(self, df: pd.DataFrame, y, factor, extra_factor=None):
+        def cal_ssw(df: pd.DataFrame, y):
+            length = df.shape[0]
+            if length==1:
+                strataVar = 0
+                lamda_1st = np.square(df[y].values[0])
+                lamda_2nd = df[y].values[0]
+            else:
+                strataVar = (length-1) * df[y].var(ddof=1)
+
+                lamda_1st = np.square(df[y].values.mean())
+                lamda_2nd = np.sqrt(length) * df[y].values.mean()
+            return strataVar, lamda_1st, lamda_2nd
+        if extra_factor==None:
+            df2 = df[[y, factor]].groupby(factor).apply(cal_ssw, y=y)
+        else:
+            df2 = df[[y]+list(set([factor, extra_factor]))].groupby([factor, extra_factor]).apply(cal_ssw, y=y)
+        df2 = df2.apply(pd.Series)
+        df2 = df2.sum()
+        strataVarSum, lamda_1st_sum, lamda_2nd_sum = df2.values
+        return strataVarSum, lamda_1st_sum, lamda_2nd_sum
+
+    @classmethod
+    def _cal_q(self, df, y, factor, extra_factor=None):
+        strataVarSum, lamda_1st_sum, lamda_2nd_sum = self._cal_ssw(df, y, factor, extra_factor)
+        TotalVar = (df.shape[0]-1)*df[y].var(ddof=1)
+        q = 1 - strataVarSum/TotalVar
+        return q, lamda_1st_sum, lamda_2nd_sum
+
+    def factor_dector(self):
+        self.factor_df = pd.DataFrame(index=["q statistic", "p value"], columns=self.factors, dtype="float32")
+        N_var = self.df[self.y].var(ddof=1)
+        N_popu = self.df.shape[0]
+        for factor in self.factors:
+            N_stra = self.df[factor].unique().shape[0]
+            q, lamda_1st_sum, lamda_2nd_sum = self._cal_q(self.df, self.y, factor)
+
+            #lamda value
+            lamda = (lamda_1st_sum - np.square(lamda_2nd_sum) / N_popu) / N_var
+            # F value
+            F_value = (N_popu - N_stra)* q / ((N_stra - 1)* (1 - q))
+            #p value
+            p_value = ncf.sf(F_value, N_stra - 1, N_popu - N_stra, nc=lamda)
+
+            self.factor_df.loc["q statistic", factor] = q
+            self.factor_df.loc["p value", factor] = p_value
+        return self.factor_df
+    
+    @classmethod
+    def _interaction_relationship(self, df):
+        out_df = pd.DataFrame(index=df.index, columns=df.columns)
+        length = len(df.index)
+        for i in range(length):
+            for j in range(i+1, length):
+                factor1, factor2 = df.index[i], df.index[j]
+                i_q = df.loc[factor2, factor1]
+                q1 = df.loc[factor1, factor1]
+                q2 = df.loc[factor2, factor2]
+
+                if (i_q <= q1 and i_q <= q2):
+                    outputRls = "Weaken, nonlinear"
+                if (i_q < max(q1, q2) and i_q > min(q1, q2)):
+                    outputRls = "Weaken, uni-"
+                if (i_q == (q1 + q2)):
+                    outputRls = "Independent"
+                if (i_q > max(q1, q2)):
+                    outputRls = "Enhance, bi-"
+                if (i_q > (q1 + q2)):
+                    outputRls = "Enhance, nonlinear"
+
+                out_df.loc[factor2, factor1] = outputRls
+        return out_df
+
+    def interaction_detector(self, relationship=False):
+        self.interaction_df = pd.DataFrame(index=self.factors, columns=self.factors, dtype="float32")
+        length = len(self.factors)
+        for i in range(0, length):
+            for j in range(0, i+1):
+                q, _, _ = self._cal_q(self.df, self.y, self.factors[i], self.factors[j])
+                self.interaction_df.loc[self.factors[i], self.factors[j]] = q
+
+        if relationship:
+            self.interaction_relationship_df = self._interaction_relationship(self.interaction_df)
+            return self.interaction_df, self.interaction_relationship_df
+        return self.interaction_df
+    
+    def ecological_detector(self):
+        self.ecological_df = pd.DataFrame(index=self.factors, columns=self.factors, dtype="float32")
+        length = len(self.factors)
+        for i in range(1, length):
+            ssw1, _, _ = self._cal_ssw(self.df, self.y, self.factors[i])
+            dfn = self.df[self.factors[i]].notna().sum()-1
+            for j in range(0, i):
+                ssw2, _, _ = self._cal_ssw(self.df, self.y, self.factors[j])
+                dfd = self.df[self.factors[j]].notna().sum()-1
+                fval = (dfn*(dfd-1)*ssw1)/(dfd*(dfn-1)*ssw2)
+                if fval<f.ppf(0.05, dfn, dfn):
+                    self.ecological_df.loc[self.factors[i], self.factors[j]] = 'Y'
+                else:
+                    self.ecological_df.loc[self.factors[i], self.factors[j]] = 'N'
+        return self.ecological_df
+
+    def plot(self, tick_fontsize=10, value_fontsize=10, colorbar_fontsize=10, show=True):
+        if isinstance(self.interaction_df, type(None)):
+            self.interaction_detector()
+        if isinstance(self.ecological_df, type(None)):
+            self.ecological_detector()
+
+        fig, ax = plt.subplots(constrained_layout=True)
+
+        im = ax.imshow(self.interaction_df.values, cmap="YlGnBu", vmin=0, vmax=1)
+        _plot_value(ax, self.interaction_df, self.ecological_df, value_fontsize=value_fontsize)
+
+        ax.set_xticks(np.arange(len(self.factors)))
+        ax.set_yticks(np.arange(len(self.factors)))
+        ax.spines['top'].set_visible(False)
+        ax.spines['right'].set_visible(False)
+
+        ax.set_xticklabels(self.factors, fontsize=tick_fontsize)
+        ax.set_yticklabels(self.factors, rotation=45, fontsize=tick_fontsize)
+        ax.tick_params(axis='y', pad=0.1)
+
+        colorbar = fig.colorbar(im, ax=ax, shrink=0.9, pad=0.01, aspect=25, extend="both")
+        colorbar.ax.tick_params(labelsize=colorbar_fontsize)
+
+        if show:
+            plt.show()
+            return ax
+        else:
+            return ax

py_geodetector-0.1.1.dist-info/METADATA

@@ -0,0 +1,63 @@
+Metadata-Version: 2.3
+Name: py_geodetector
+Version: 0.1.1
+Summary: A simple Python package for the geodetector
+Project-URL: Homepage, https://github.com/djw-easy/GeoDetector
+Project-URL: Bug Tracker, https://github.com/djw-easy/GeoDetector/issues
+Author-email: djw <djweasy@163.com>
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Requires-Python: >=3.7
+Description-Content-Type: text/markdown
+
+# A simple Python package for the geodetector 
+
+# Install
+
+```
+pip install py-geodetector
+```
+
+# Usage
+
+A quick example of geodetector usage is given in the ./example.ipynb.
+
+```python
+from py_geodetector import load_example_data, GeoDetector
+
+# load example data
+df = load_example_data()
+
+gd = GeoDetector(df)
+# factor detect
+factor_df = gd.factor_dector()
+
+# interaction detect
+interaction_df = gd.interaction_detector()
+# or you can generate the interaction relationship as the same time
+interaction_df, interaction_relationship_df = gd.interaction_detector(relationship=True)
+
+# ecological detect
+ecological_df = gd.ecological_detector()
+
+# plot 
+# use a heatmap visualize the interaction detect result, 
+# red text means that the ecological detection results show a significant difference
+gd.plot(value_fontsize=14, tick_fontsize=16, colorbar_fontsize=14);
+```
+
+# Reference
+
+```
+@article{wang2010geographical,
+  title={Geographical detectors-based health risk assessment and its application in the neural tube defects study of the Heshun Region, China},
+  author={Wang, Jin-Feng and Li, Xin-Hu and Christakos, George and Liao, Yi-Lan and Zhang, Tin and Gu, Xue and Zheng, Xiao-Ying},
+  journal={International Journal of Geographical Information Science},
+  volume={24},
+  number={1},
+  pages={107-127},
+  year={2010},
+  publisher={Taylor \& Francis}
+}
+```

py_geodetector-0.1.1.dist-info/RECORD

@@ -0,0 +1,6 @@
+py_geodetector/__init__.py,sha256=bhyJq1ipXNRhG-zPbF_5YUbJ_sana5rJ-FG-ogBNh_U,57
+py_geodetector/geodetector.py,sha256=R-3OlNrgghwZTdieIbhB5J3N3nl2-ix6bjwjyVmh1E8,8240
+py_geodetector/example_data/disease.csv,sha256=sodkE21Xw-eSlGWxLE1by7hSHQk2FjNxb4ncWdyWLmE,2231
+py_geodetector-0.1.1.dist-info/METADATA,sha256=ILLj1V2F_SoPoiJCdaAjmRCT_W4gSeLqCeKAmoznUL0,1853
+py_geodetector-0.1.1.dist-info/WHEEL,sha256=1yFddiXMmvYK7QYTqtRNtX66WJ0Mz8PYEiEUoOUUxRY,87
+py_geodetector-0.1.1.dist-info/RECORD,,

py_geodetector-0.1.1.dist-info/WHEEL

@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: hatchling 1.25.0
+Root-Is-Purelib: true
+Tag: py3-none-any