diffindiff 2.0.2__tar.gz → 2.0.4__tar.gz

{diffindiff-2.0.2 → diffindiff-2.0.4}/PKG-INFO

@@ -1,10 +1,24 @@
-Metadata-Version: 2.1
+Metadata-Version: 2.4
 Name: diffindiff
-Version: 2.0.2
+Version: 2.0.4
 Summary: diffindiff: Python library for convenient Difference-in-Differences Analyses
 Author: Thomas Wieland
 Author-email: geowieland@googlemail.com
 Description-Content-Type: text/markdown
+Requires-Dist: numpy
+Requires-Dist: pandas
+Requires-Dist: statsmodels
+Requires-Dist: matplotlib
+Requires-Dist: datetime
+Requires-Dist: scikit-learn
+Requires-Dist: xgboost
+Requires-Dist: lightgbm
+Dynamic: author
+Dynamic: author-email
+Dynamic: description
+Dynamic: description-content-type
+Dynamic: requires-dist
+Dynamic: summary
 
 # diffindiff: Difference-in-Differences (DiD) Analysis Python Library
 
@@ -24,17 +38,18 @@ Thomas Wieland [ORCID](https://orcid.org/0000-0001-5168-9846) [EMail](mailto:geo
   - Create predictive counterfactuals
 - **DiD analysis**: 
   - Perfom standard DiD analysis  
-  - Model Extensions:
+  - Model extensions:
     - Staggered adoption
     - Multiple treatments
     - Two-way fixed effects models
     - Group- or individual-specific treatment effects
     - Group- or individual-specific time trends
     - Including covariates
-    - After-treatment period
+    - Including fter-treatment period
     - Triple Difference (DDD)
     - Own counterfactuals
-    - Bonferroni correction
+    - Bonferroni correction for treatment effects
+    - Placebo test
 - **Visualization**:
   - Plot observed and expected time course of treatment and control group
   - Plot expected time course of treatment group and counterfactual
@@ -46,7 +61,6 @@ Thomas Wieland [ORCID](https://orcid.org/0000-0001-5168-9846) [EMail](mailto:geo
   - Test for type of adoption
   - Test whether the panel dataset is balanced
   - Test for parallel trend assumption
-  - Placebo test
 
 
 ## Literature

{diffindiff-2.0.2 → diffindiff-2.0.4}/README.md

@@ -16,17 +16,18 @@ Thomas Wieland [ORCID](https://orcid.org/0000-0001-5168-9846) [EMail](mailto:geo
   - Create predictive counterfactuals
 - **DiD analysis**: 
   - Perfom standard DiD analysis  
-  - Model Extensions:
+  - Model extensions:
     - Staggered adoption
     - Multiple treatments
     - Two-way fixed effects models
     - Group- or individual-specific treatment effects
     - Group- or individual-specific time trends
     - Including covariates
-    - After-treatment period
+    - Including fter-treatment period
     - Triple Difference (DDD)
     - Own counterfactuals
-    - Bonferroni correction
+    - Bonferroni correction for treatment effects
+    - Placebo test
 - **Visualization**:
   - Plot observed and expected time course of treatment and control group
   - Plot expected time course of treatment group and counterfactual
@@ -38,7 +39,6 @@ Thomas Wieland [ORCID](https://orcid.org/0000-0001-5168-9846) [EMail](mailto:geo
   - Test for type of adoption
   - Test whether the panel dataset is balanced
   - Test for parallel trend assumption
-  - Placebo test
 
 
 ## Literature

{diffindiff-2.0.2 → diffindiff-2.0.4}/diffindiff/didanalysis.py

@@ -1,11 +1,13 @@
-#-------------------------------------------------------------------------------
-# Name:        didanalysis (diffindiff)
+#-----------------------------------------------------------------------
+# Name:        didanalysis (diffindiff package)
 # Purpose:     Analysis functions for difference-in-differences analyses
-# Author:      Thomas Wieland (mail: geowieland@googlemail.com, ORCID: 0000-0001-5168-9846)
-# Version:     2.0.2
-# Last update: 2025-04-16 17:10
+# Author:      Thomas Wieland 
+#              ORCID: 0000-0001-5168-9846
+#              mail: geowieland@googlemail.com              
+# Version:     2.0.4
+# Last update: 2025-04-18 15:55
 # Copyright (c) 2025 Thomas Wieland
-#-------------------------------------------------------------------------------
+#-----------------------------------------------------------------------
 
 
 import pandas as pd
@@ -163,11 +165,11 @@ class DiffModel:
             treatment_diagnostics_rows.append({
                 "Treatment": value["treatment"],
                 "Type of adoption": adoption_type,
-                "No-treatment control group": no_treatment,
-                "Treatment group (N)": treatment_group_size,
-                "Control group (N)": control_group_size,
+                "No-treatment control group": no_treatment,                
                 "Parallel trends (pre)": is_parallel,
-                "Format": value["treatment_format"]
+                "Format": value["treatment_format"],
+                "Treatment group (N)": treatment_group_size,
+                "Control group (N)": control_group_size
             })
 
             if no_treatment == "NO" and adoption_type == "Simultaneous":
@@ -366,7 +368,7 @@ class DiffModel:
 
             for key, value in covariates_effects.items():
                 covariates_effects_rows.append({
-                    "Covariates": value["Coefficient"],
+                    "": value["Coefficient"],
                     "Estimate": value["Estimate"],
                     "SE": value["SE"],
                     "t": value["t"],
@@ -525,8 +527,9 @@ class DiffModel:
             covariates_effects_df["CI lower"] = covariates_effects_df["CI lower"].map(lambda x: f"{x:,.3f}")
             covariates_effects_df["CI upper"] = covariates_effects_df["CI upper"].map(lambda x: f"{x:,.3f}")
             covariates_effects_df.iloc[:, 0] = covariates_effects_df.iloc[:, 0].apply(lambda x: f"{x:<{max_width_column1}}")
+            print("Covariates") 
             print(covariates_effects_df.to_string(index=False))
-        if not show_covariates:
+        if not show_covariates or no_covariates == 0:
             if no_covariates > 0:
                 print ("Covariates                 YES")
             else:
@@ -566,6 +569,13 @@ class DiffModel:
             index = treatment_diagnostics_df.columns)
         treatment_diagnostics_df_t = treatment_diagnostics_df_t.iloc[1:]
         print(treatment_diagnostics_df_t)
+        if model_config["no_treatments"] > 1:
+            untreated = diffindiff.didtools.untreated_units(
+                data = model_data,
+                unit_col = model_config["unit_col"],
+                treatment_col = model_config["treatment_col"]
+                )
+            print ("Units with >=1 treatment(s): " + str(untreated[0]) + ", non-treated units: " + str(untreated[1]))   
 
         print("-" * total_width)
         print ("Input data diagnostics")

{diffindiff-2.0.2 → diffindiff-2.0.4}/diffindiff/diddata.py

@@ -1,11 +1,14 @@
-#-------------------------------------------------------------------------------
-# Name:        diddata (diffindiff)
+#-----------------------------------------------------------------------
+# Name:        diddata (diffindiff package)
 # Purpose:     Creating data for Difference-in-Differences Analysis
-# Author:      Thomas Wieland (mail: geowieland@googlemail.com, ORCID: 0000-0001-5168-9846)
-# Version:     2.0.2
-# Last update: 2025-04-16 17:10
+# Author:      Thomas Wieland 
+#              ORCID: 0000-0001-5168-9846
+#              mail: geowieland@googlemail.com              
+# Version:     2.0.4
+# Last update: 2025-04-18 15:38
 # Copyright (c) 2025 Thomas Wieland
-#-------------------------------------------------------------------------------
+#-----------------------------------------------------------------------
+
 
 import pandas as pd
 import numpy as np

{diffindiff-2.0.2 → diffindiff-2.0.4}/diffindiff/didtools.py

@@ -1,11 +1,13 @@
-# -------------------------------------------------------------------------------
-# Name:        didtools (diffindiff)
-# Purpose:     Creating data for Difference-in-Differences Analysis
-# Author:      Thomas Wieland (mail: geowieland@googlemail.com, ORCID: 0000-0001-5168-9846)
-# Version:     2.0.2
-# Last update: 2025-04-16 17:10
+#-----------------------------------------------------------------------
+# Name:        didtools (diffindiff package)
+# Purpose:     Additional tools for Difference-in-Differences Analysis
+# Author:      Thomas Wieland 
+#              ORCID: 0000-0001-5168-9846
+#              mail: geowieland@googlemail.com              
+# Version:     2.0.4
+# Last update: 2025-04-18 15:38
 # Copyright (c) 2025 Thomas Wieland
-#-------------------------------------------------------------------------------
+#-----------------------------------------------------------------------
 
 
 import pandas as pd
@@ -34,11 +36,11 @@ def check_columns(
         raise ValueError(f"Data do not contain column(s): {', '.join(missing_columns)}")
 
 def is_balanced(
-    data,
-    unit_col,
-    time_col,
-    outcome_col,
-    other_cols = None
+    data: pd.DataFrame,
+    unit_col: str,
+    time_col: str,
+    outcome_col: str,
+    other_cols: list = None
     ):
 
     unit_freq = data[unit_col].nunique()
@@ -58,8 +60,8 @@ def is_balanced(
         return True
 
 def is_binary(
-    data,
-    treatment_col
+    data: pd.DataFrame,
+    treatment_col: str
     ):
     
     unique_values = set(data[treatment_col].dropna().unique())
@@ -76,7 +78,7 @@ def is_binary(
         return [False, "Unknown"]             
 
 def is_missing(
-    data,
+    data: pd.DataFrame,
     drop_missing: bool = True,
     missing_replace_by_zero: bool = False
     ):
@@ -104,10 +106,10 @@ def is_missing(
         ]
 
 def is_simultaneous(
-    data,
-    unit_col,
-    time_col,
-    treatment_col,
+    data: pd.DataFrame,
+    unit_col: str,
+    time_col: str,
+    treatment_col: str,
     pre_post = False
     ):
 
@@ -125,9 +127,9 @@ def is_simultaneous(
     return col_identical
 
 def is_notreatment(
-    data,
-    unit_col,
-    treatment_col
+    data: pd.DataFrame,
+    unit_col: str,
+    treatment_col: str
     ):
 
     data_relevant = data[[unit_col, treatment_col]]
@@ -150,12 +152,63 @@ def is_notreatment(
         control_group
         ]
 
+def treatment_group_col(
+    data: pd.DataFrame,
+    unit_col: str,
+    treatment_col: str,
+    create_TG_col: str = "TG"   
+    ):
+    
+    isnotreatment = is_notreatment(
+        data = data,
+        unit_col = unit_col,
+        treatment_col = treatment_col
+        )
+    
+    if not isnotreatment[0]:
+        print ("Model data does not contain a no-treatment control group. Treatment group column is constant = 1.")
+    
+    if create_TG_col in data.columns:
+        create_TG_col = "TG_"+treatment_col
+        print ("Column " + create_TG_col + " already exists. Saving treatment group in column TG_" + treatment_col)
+
+    treatment_group = isnotreatment[1]
+
+    data[create_TG_col] = 0
+    data.loc[data[unit_col].astype(str).isin(treatment_group), create_TG_col] = 1
+
+    return [
+        data, 
+        isnotreatment[0], 
+        create_TG_col
+        ]
+
+def untreated_units(
+    data: pd.DataFrame,
+    unit_col: str,
+    treatment_col: list   
+    ):
+
+    unit_sum = data.groupby(unit_col)[treatment_col].sum().sum(axis=1).reset_index(name="sum")
+
+    units_treated = unit_sum.loc[unit_sum["sum"] > 0, unit_col]
+    units_nontreated = unit_sum.loc[unit_sum["sum"] == 0, unit_col]
+    no_units_treated = len(units_treated)
+    no_units_nontreated = len(units_nontreated)
+
+    return [
+        no_units_treated,
+        no_units_nontreated,
+        units_treated,
+        units_nontreated
+        ]
+
 def is_parallel(
-    data,
-    unit_col,
-    time_col,
-    treatment_col,
-    outcome_col,
+    data: pd.DataFrame,
+    unit_col: str,
+    time_col: str,
+    treatment_col: str,
+    outcome_col: str,
     pre_post = False,
     alpha = 0.05
     ):
@@ -206,10 +259,10 @@ def is_parallel(
         ]
 
 def date_counter(
-        df,
-        date_col, 
-        new_col = "date_counter"
-        ):
+    df: pd.DataFrame,
+    date_col: str, 
+    new_col: str = "date_counter"
+    ):
     
     dates = df[date_col].unique()
 
@@ -226,6 +279,7 @@ def date_counter(
     return df
 
 def unique(data):
+
     if data is None or (isinstance(data, (list, np.ndarray, pd.Series, pd.DataFrame)) and len(data) == 0):
         return []
     
@@ -269,8 +323,9 @@ def model_wrapper(
     lgbm_learning_rate = 0.1,
     random_state = 71
     ):
-    if model_type not in ["ols", "olsbg", "dtbg", "rf", "gb", "knn", "svr", "xgb", "lgbm", "catboost"]:
-        raise ValueError("Please enter a valid model type")
+
+    if model_type not in ["ols", "olsbg", "dtbg", "rf", "gb", "knn", "svr", "xgb", "lgbm"]:
+        raise ValueError("Please enter a valid model type ('ols', 'olsbg', 'dtbg', 'rf', 'gb', 'knn', 'svr', 'xgb', 'lgbm')")
     
     X_train, X_test, y_train, y_test = train_test_split(
         X, 
@@ -348,10 +403,10 @@ def model_wrapper(
         ]
     
 def treatment_times(
-    data,
-    unit_col,
-    time_col,
-    treatment_col
+    data: pd.DataFrame,
+    unit_col: str,
+    time_col: str,
+    treatment_col: str
     ):
     
     check_columns(
@@ -389,10 +444,10 @@ def clean_column_name(value):
     return value.strip('_')
     
 def to_dummies(
-    data, 
-    col,
-    drop_first = False,
-    prefix = "DUMMY"
+    data: pd.DataFrame, 
+    col: str,
+    drop_first: bool = False,
+    prefix: str = "DUMMY"
     ):     
     
     unique_values = data[col].astype(str).unique()

{diffindiff-2.0.2 → diffindiff-2.0.4}/diffindiff/tests/tests_diffindiff.py

@@ -1,16 +1,20 @@
-#------------------------------------------------------------------------------------------
-# Name:        tests_diffindiff
+#-----------------------------------------------------------------------
+# Name:        tests_diffindiff (diffindiff package)
 # Purpose:     Tests and examples for the diffindiff package
-# Author:      Thomas Wieland (mail: geowieland@googlemail.com, ORCID: 0000-0001-5168-9846)
-# Version:     2.0.2
-# Last update: 2025-04-16 17:10
+# Author:      Thomas Wieland 
+#              ORCID: 0000-0001-5168-9846
+#              mail: geowieland@googlemail.com              
+# Version:     2.0.4
+# Last update: 2025-04-18 15:38
 # Copyright (c) 2025 Thomas Wieland
-#------------------------------------------------------------------------------------------
+#-----------------------------------------------------------------------
 
 
 import pandas as pd
 from diffindiff.didanalysis import DiffModel, did_analysis
 from diffindiff.diddata import DiffGroups, create_groups, DiffTreatment, create_treatment, DiffData, merge_data, create_data
+from diffindiff.didtools import untreated_units
+
 
 # Example 1: Effect of a curfew in German counties in the first
 # wave of the COVID-19 pandemic (DiD pre-post analysis)

{diffindiff-2.0.2 → diffindiff-2.0.4}/diffindiff.egg-info/PKG-INFO

@@ -1,10 +1,24 @@
-Metadata-Version: 2.1
+Metadata-Version: 2.4
 Name: diffindiff
-Version: 2.0.2
+Version: 2.0.4
 Summary: diffindiff: Python library for convenient Difference-in-Differences Analyses
 Author: Thomas Wieland
 Author-email: geowieland@googlemail.com
 Description-Content-Type: text/markdown
+Requires-Dist: numpy
+Requires-Dist: pandas
+Requires-Dist: statsmodels
+Requires-Dist: matplotlib
+Requires-Dist: datetime
+Requires-Dist: scikit-learn
+Requires-Dist: xgboost
+Requires-Dist: lightgbm
+Dynamic: author
+Dynamic: author-email
+Dynamic: description
+Dynamic: description-content-type
+Dynamic: requires-dist
+Dynamic: summary
 
 # diffindiff: Difference-in-Differences (DiD) Analysis Python Library
 
@@ -24,17 +38,18 @@ Thomas Wieland [ORCID](https://orcid.org/0000-0001-5168-9846) [EMail](mailto:geo
   - Create predictive counterfactuals
 - **DiD analysis**: 
   - Perfom standard DiD analysis  
-  - Model Extensions:
+  - Model extensions:
     - Staggered adoption
     - Multiple treatments
     - Two-way fixed effects models
     - Group- or individual-specific treatment effects
     - Group- or individual-specific time trends
     - Including covariates
-    - After-treatment period
+    - Including fter-treatment period
     - Triple Difference (DDD)
     - Own counterfactuals
-    - Bonferroni correction
+    - Bonferroni correction for treatment effects
+    - Placebo test
 - **Visualization**:
   - Plot observed and expected time course of treatment and control group
   - Plot expected time course of treatment group and counterfactual
@@ -46,7 +61,6 @@ Thomas Wieland [ORCID](https://orcid.org/0000-0001-5168-9846) [EMail](mailto:geo
   - Test for type of adoption
   - Test whether the panel dataset is balanced
   - Test for parallel trend assumption
-  - Placebo test
 
 
 ## Literature

{diffindiff-2.0.2 → diffindiff-2.0.4}/setup.py

@@ -7,7 +7,7 @@ def read_README():
     
 setup(
     name='diffindiff',
-    version='2.0.2',
+    version='2.0.4',
     description='diffindiff: Python library for convenient Difference-in-Differences Analyses',
     packages=find_packages(include=["diffindiff", "diffindiff.tests"]),
     include_package_data=True,