geocif 0.1.70__tar.gz → 0.1.72__tar.gz

{geocif-0.1.70/geocif.egg-info → geocif-0.1.72}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: geocif
-Version: 0.1.70
+Version: 0.1.72
 Summary: Models to visualize and forecast crop conditions and yields
 Home-page: https://ritviksahajpal.github.io/yield_forecasting/
 Author: Ritvik Sahajpal
@@ -24,6 +24,10 @@ License-File: LICENSE
 [![image](https://img.shields.io/conda/vn/conda-forge/geocif.svg)](https://anaconda.org/conda-forge/geocif)
 
 
+**Generate Climatic Impact-Drivers (CIDs) from Earth Observation (EO) data**
+
+[Climatic Impact-Drivers for Crop Yield Assessment at NASA Harvest](https://www.loom.com/share/5c2dc62356c6406193cd9d9725c2a6a9)
+
 **Models to visualize and forecast crop conditions and yields**
 
 

{geocif-0.1.70 → geocif-0.1.72}/README.md

@@ -5,6 +5,10 @@
 [![image](https://img.shields.io/conda/vn/conda-forge/geocif.svg)](https://anaconda.org/conda-forge/geocif)
 
 
+**Generate Climatic Impact-Drivers (CIDs) from Earth Observation (EO) data**
+
+[Climatic Impact-Drivers for Crop Yield Assessment at NASA Harvest](https://www.loom.com/share/5c2dc62356c6406193cd9d9725c2a6a9)
+
 **Models to visualize and forecast crop conditions and yields**
 
 

{geocif-0.1.70 → geocif-0.1.72}/geocif/analysis.py

@@ -560,12 +560,12 @@ class Geoanalysis:
         )
 
         # --- For computing median yields ---
-        # Compute median yield for 2014 - 2018
-        df_median_2014_2018 = (
-            df_all[df_all["Harvest Year"].between(2014, 2018)]
+        # Compute median yield for 2018 - 2022
+        df_median_2018_2022 = (
+            df_all[df_all["Harvest Year"].between(2018, 2022)]
             .groupby("Region")["Yield (tn per ha)"]
-            .median()
-            .rename(f"Median Yield (tn per ha) (2014-2018)")
+            .mean()
+            .rename(f"Median Yield (tn per ha) (2018-2022)")
             .reset_index()
         )
 
@@ -573,7 +573,7 @@ class Geoanalysis:
         df_median_2013_2017 = (
             df_all[df_all["Harvest Year"].between(2013, 2017)]
             .groupby("Region")["Yield (tn per ha)"]
-            .median()
+            .mean()
             .rename("Median Yield (tn per ha) (2013-2017)")
             .reset_index()
         )
@@ -581,7 +581,7 @@ class Geoanalysis:
         # Merge the median yield columns with the % of total production dataframe
         df_historic = (
             df_pct
-            .merge(df_median_2014_2018, on="Region", how="left")
+            .merge(df_median_2018_2022, on="Region", how="left")
             .merge(df_median_2013_2017, on="Region", how="left")
         )
 
@@ -885,7 +885,9 @@ class Geoanalysis:
         # Get the ML section
         df_ml = self.df_config[self.df_config["Section"] == "ML"]
 
-        self.countries = ["malawi"]
+        self.countries = ast.literal_eval(
+            df_ml[df_ml["Option"] == "countries"]["Value"].values[0]
+        )
         for country in self.countries:
             df = self.df_config[self.df_config["Section"] == country]
 

{geocif-0.1.70 → geocif-0.1.72}/geocif/indices_runner_angola.py

@@ -174,7 +174,7 @@ class cei_runner(base.BaseGeo):
         combinations = [i for i in combinations if f"{country}_maize_s1" in i[3]]
 
         if True:
-            num_cpu = int(cpu_count() * 0.9)
+            num_cpu = int(cpu_count() * 0.2)
             with Pool(num_cpu) as p:
                 for i, _ in enumerate(p.imap_unordered(indices.process, combinations)):
                     pass

{geocif-0.1.70 → geocif-0.1.72}/geocif/indices_runner_madagascar.py

@@ -174,7 +174,7 @@ class cei_runner(base.BaseGeo):
         combinations = [i for i in combinations if f"{country}_maize_s1" in i[3]]
 
         if True:
-            num_cpu = int(cpu_count() * 0.1)
+            num_cpu = int(cpu_count() * 0.2)
             with Pool(num_cpu) as p:
                 for i, _ in enumerate(p.imap_unordered(indices.process, combinations)):
                     pass

{geocif-0.1.70 → geocif-0.1.72}/geocif/indices_runner_malawi.py

@@ -174,7 +174,7 @@ class cei_runner(base.BaseGeo):
         combinations = [i for i in combinations if f"{country}_maize_s1" in i[3]]
 
         if True:
-            num_cpu = int(cpu_count() * 0.1)
+            num_cpu = int(cpu_count() * 0.2)
             with Pool(num_cpu) as p:
                 for i, _ in enumerate(p.imap_unordered(indices.process, combinations)):
                     pass

{geocif-0.1.70 → geocif-0.1.72}/geocif/indices_runner_mozambique.py

@@ -174,7 +174,7 @@ class cei_runner(base.BaseGeo):
         combinations = [i for i in combinations if f"{country}_maize_s1" in i[3]]
 
         if True:
-            num_cpu = int(cpu_count() * 0.1)
+            num_cpu = int(cpu_count() * 0.2)
             with Pool(num_cpu) as p:
                 for i, _ in enumerate(p.imap_unordered(indices.process, combinations)):
                     pass

{geocif-0.1.70 → geocif-0.1.72}/geocif/indices_runner_south_africa.py

@@ -174,7 +174,7 @@ class cei_runner(base.BaseGeo):
         combinations = [i for i in combinations if f"{country}_maize_s1" in i[3]]
 
         if True:
-            num_cpu = int(cpu_count() * 0.8)
+            num_cpu = int(cpu_count() * 0.2)
             with Pool(num_cpu) as p:
                 for i, _ in enumerate(p.imap_unordered(indices.process, combinations)):
                     pass

{geocif-0.1.70 → geocif-0.1.72}/geocif/indices_runner_zambia.py

@@ -174,7 +174,7 @@ class cei_runner(base.BaseGeo):
         combinations = [i for i in combinations if f"{country}_maize_s1" in i[3]]
 
         if True:
-            num_cpu = int(cpu_count() * 0.1)
+            num_cpu = int(cpu_count() * 0.2)
             with Pool(num_cpu) as p:
                 for i, _ in enumerate(p.imap_unordered(indices.process, combinations)):
                     pass

{geocif-0.1.70 → geocif-0.1.72}/geocif/indices_runner_zimbabwe.py

@@ -174,7 +174,7 @@ class cei_runner(base.BaseGeo):
         combinations = [i for i in combinations if f"{country}_maize_s1" in i[3]]
 
         if True:
-            num_cpu = int(cpu_count() * 0.1)
+            num_cpu = int(cpu_count() * 0.2)
             with Pool(num_cpu) as p:
                 for i, _ in enumerate(p.imap_unordered(indices.process, combinations)):
                     pass

{geocif-0.1.70 → geocif-0.1.72}/geocif/playground/aa.py

@@ -1,3 +1,89 @@
+import geopandas as gpd
+import pandas as pd
+import matplotlib.pyplot as plt
+import matplotlib as mpl
+
+# --- 1. Read data ---
+dg = gpd.read_file(r"D:\Users\ritvik\projects\GEOGLAM\safrica.shp")
+df = pd.read_csv(r"D:\Users\ritvik\projects\GEOGLAM\geocif_march_2025.csv")
+
+# --- 2. Create the new "Country Region" column ---
+dg['Country Region'] = (
+    dg.apply(
+        lambda row: (
+            f"{row['ADMIN0']} {row['ADMIN2']}"
+            if pd.notnull(row['ADMIN2'])
+            else f"{row['ADMIN0']} {row['ADMIN1']}"
+        ),
+        axis=1
+    )
+    .str.lower()
+    .str.replace(' ', '_')
+)
+
+# --- 3. Merge shapefile with CSV ---
+merged = dg.merge(df, left_on='Country Region', right_on='Country Region', how='right')
+
+# --- 4. Rename columns ---
+merged.rename(
+    columns={
+        '% Anomaly (2013-2017)': '2013_2017',
+        '% Anomaly (2018-2022)': '2018_2022'
+    },
+    inplace=True
+)
+
+# Optional: Write out merged shapefile
+merged.to_file(r"D:\Users\ritvik\projects\GEOGLAM\safrica_geocif_march_2025.shp")
+
+# --- 5. Plot ---
+fig, ax = plt.subplots(1, 2, figsize=(20, 10))
+
+# Reduce horizontal space between subplots
+plt.subplots_adjust(wspace=0.05)
+
+# Shared color normalization
+norm = mpl.colors.Normalize(vmin=-40, vmax=40)
+
+# Plot the anomaly maps (no country boundaries)
+merged.plot(
+    column='2013_2017',
+    cmap='BrBG',
+    norm=norm,
+    ax=ax[0],
+    legend=False
+)
+ax[0].set_title('Maize Yield Forecast % Anomaly (2013-2017)')
+ax[0].axis('off')
+
+merged.plot(
+    column='2018_2022',
+    cmap='BrBG',
+    norm=norm,
+    ax=ax[1],
+    legend=False
+)
+ax[1].set_title('Maize Yield Forecast % Anomaly (2018-2022)')
+ax[1].axis('off')
+
+# Create a single horizontal colorbar
+sm = mpl.cm.ScalarMappable(norm=norm, cmap='BrBG')
+sm.set_array([])
+cbar = fig.colorbar(
+    sm,
+    ax=ax.ravel().tolist(),
+    orientation='horizontal',
+    fraction=0.05,
+    pad=0.05,
+    extend='both'
+)
+cbar.set_label('% Anomaly')
+
+plt.savefig(r"D:\Users\ritvik\projects\GEOGLAM\maize_yield_forecast_anomaly.png", dpi=300)
+
+
+
+breakpoint()
 from great_tables import GT, html
 import pandas as pd
 

{geocif-0.1.70 → geocif-0.1.72}/geocif/viz/tmp.py

@@ -1,13 +1,15 @@
 import geopandas as gpd
 import palettable as pal
 import matplotlib.colors as mcolors
-
+import matplotlib.pyplot as plt
+import seaborn as sns
+import numpy as np
 import pandas as pd
 import glob
 import os
 
 # 1. Specify the directory containing your .dta files:
-data_dir =  r"C:\Users\ritvik\Downloads\maize_yield (2)\maize_yield"
+data_dir =  r"C:\Users\ritvik\Downloads\maize_yield\maize_yield"
 
 # 2. Use glob to find all .dta files in that directory:
 dta_files = glob.glob(os.path.join(data_dir, "*.dta"))
@@ -41,7 +43,7 @@ merged_df['W_CODE'] = '7' + merged_df['W_CODE']
 merged_df['W_CODE'] = merged_df['W_CODE'].str.replace('.0', '')
 merged_df['W_CODE'] = merged_df['W_CODE'].astype(int)
 
-dg = gpd.read_file(r"wolayita_dissolved.shp")
+dg = gpd.read_file(r"D:\Users\ritvik\projects\GEOGLAM\Input\Global_Datasets\Regions\Shps\wolayita_dissolved.shp")
 dg = dg[['W_CODE', 'W_NAME']]
 
 # Merge the two dataframes on W_CODE
@@ -52,10 +54,201 @@ merged_df = merged_df.dropna(subset=['PROD98CQ', 'AREAH'])
 
 # Compte yield column
 merged_df['yield'] = merged_df['PROD98CQ'] / merged_df['AREAH']
+merged_df.to_csv(r'D:\Users\ritvik\projects\GEOGLAM\Output\crop_condition\March_27_2025\plots\EWCM\kabele.csv', index=False)
+breakpoint()
+# Add a histogram showing distribution of yields, use separate line and color for each 5-year period
+# Create a new column 'Year Group' which groups years into 5-year periods
+import pandas as pd
+import seaborn as sns
+import matplotlib.pyplot as plt
+
+# 1. Group years into 5-year periods
+merged_df['Year Group'] = pd.cut(
+    merged_df['YEAR'],
+    bins=range(2005, 2026, 5),
+    right=False
+)
+
+# 2. Create a FacetGrid with one facet per Year Group
+g = sns.FacetGrid(
+    merged_df,
+    col="Year Group",
+    col_wrap=2,
+    height=4,
+    sharex=True,
+    sharey=True
+)
+
+# 3. Map an ECDF plot to each facet
+#    'skyblue' is used for consistency with your original color choice
+g.map(sns.ecdfplot, 'yield', color='skyblue')
+
+# 4. Add vertical/horizontal lines, annotations, etc. in each facet
+for ax, year_group in zip(g.axes.flatten(), g.col_names):
+    # -- Subset data for this particular facet
+    subset = merged_df[merged_df['Year Group'] == year_group]
+
+    # -- Vertical line at yield=16
+    ax.axvline(x=16, color='red', linestyle='--')
+
+    # -- Annotate the line at yield=16
+    ax.annotate(
+        '16 QQ/ha',
+        xy=(16, 0.8),  # x=16, y=0.8 in data coordinates (cumulative fraction)
+        xytext=(5, 0),  # offset the text to the right by 5 points
+        textcoords='offset points',
+        rotation=90,
+        color='red',
+        ha='center',
+        va='center',
+        fontsize=9
+    )
+
+    # -- Horizontal lines for quintiles on the y-axis (20%, 40%, 60%, 80%)
+    for q in [0.2, 0.4, 0.6, 0.8]:
+        ax.axhline(y=q, color='green', linestyle='--')
+        # (Optional) label each horizontal line:
+        # ax.text(ax.get_xlim()[1]*0.9, q, f"{int(q*100)}%",
+        #         va='center', ha='right', color='green', fontsize=9)
+
+    # -- Number of observations in top-right corner
+    n_obs = len(subset)
+    ax.text(
+        0.95, 0.95,
+        f"N = {n_obs}",
+        transform=ax.transAxes,
+        ha='right',
+        va='top',
+        fontsize=9,
+        color='black'
+    )
+
+# 5. Common title
+plt.subplots_adjust(top=0.9)
+g.fig.suptitle("Yield Distribution at Kabele Level (ECDF)", fontsize=16)
+
+plt.show()
+
+# Find the percentage of values below 16 QQ/ha and Year Group 2015-2020 or 2020-2025
+#below_16 = merged_df[(merged_df['yield'] < 16) & (merged_df['YEAR'] >=2015) & (merged_df['YEAR'] < 2025)]
+#breakpoint()
+#below_16_pct = below_16.mean() * 100
+#print(f"Percentage of yields below 16 QQ/ha at Kabele level: {below_16_pct:.2f}%")
+
+
+# Create a heatmap showing the number of unique FA's per DIST and YEAR combination
+# Group by DIST and YEAR, then count the number of unique FA's
+df_fa_counts = merged_df.groupby(['DIST', 'YEAR'])['FA'].nunique().reset_index(name='FA_Count')
+
+# Pivot the data so that rows = DIST, columns = YEAR, values = FA_Count
+df_fa_pivot = df_fa_counts.pivot(index='DIST', columns='YEAR', values='FA_Count')
+
+# Change year column to type int
+df_fa_pivot.columns = df_fa_pivot.columns.astype(int)
+
+# Create the heatmap
+
+plt.figure(figsize=(12, 8))
+sns.heatmap(
+    df_fa_pivot,
+    cmap='viridis',       # color map; try 'coolwarm' or others
+    annot=True,           # show numeric values in each cell
+    fmt="g",            # format numbers (2 decimal places)
+    linewidths=.5         # line width between cells
+)
+plt.show()
 
 # create a new dataframe which computes average yield by W_NAME for each year, do a weighted average using FWEIGHT column
 df_avg_yield = merged_df.groupby(['W_NAME', 'YEAR']).apply(lambda x: np.average(x['yield'], weights=x['FWEIGHT'])).reset_index(name='yield')
 
+# Add a histogram showing distribution of yields, use separate line and color for each 5-year period
+# Create a new column 'Year Group' which groups years into 5-year periods
+import pandas as pd
+import seaborn as sns
+import matplotlib.pyplot as plt
+
+# 1. Create custom bins and labels so that the last group is 2020-2021
+import pandas as pd
+import seaborn as sns
+import matplotlib.pyplot as plt
+
+# 1. Define custom bins so that the last group is labeled 2020–2021
+import pandas as pd
+import seaborn as sns
+import matplotlib.pyplot as plt
+
+# 1. Define custom bins so that the last group is labeled 2020–2021
+df_avg_yield['Year Group'] = pd.cut(
+    df_avg_yield['YEAR'],
+    bins=[2005, 2010, 2015, 2020, 2022],  # stops at 2022 so the label reads "2020–2021"
+    labels=['2005–2009', '2010–2014', '2015–2019', '2020–2021'],
+    right=False
+)
+
+# 2. Create a FacetGrid by Year Group
+g = sns.FacetGrid(
+    df_avg_yield,
+    col='Year Group',
+    col_wrap=2,
+    height=4,
+    sharex=True,
+    sharey=True
+)
+
+# 3. Map an ECDF plot (instead of histogram) in each facet
+g.map_dataframe(sns.ecdfplot, x='yield')
+
+# 4. Add lines and annotations to each facet
+for i, ax in enumerate(g.axes.flatten()):
+    # Subset the data for the current Year Group
+    year_group = g.col_names[i]
+    subset = df_avg_yield[df_avg_yield['Year Group'] == year_group]
+
+    # -- Vertical line at yield=16
+    ax.axvline(x=16, color='red', linestyle='--')
+    # Annotate that line near the top of the plot
+    ax.annotate(
+        "16 QQ/ha",
+        xy=(16, 0.8),  # (x=16, y=0.8 in data coords for the y-axis)
+        xytext=(5, 0),
+        textcoords='offset points',
+        rotation=90,
+        color='red',
+        ha='center',
+        va='center',
+        fontsize=9
+    )
+
+    # -- Horizontal lines at 20%, 40%, 60%, 80% (quintiles in terms of fraction)
+    for q in [0.2, 0.4, 0.6, 0.8]:
+        ax.axhline(y=q, color='green', linestyle='--')
+
+    # -- Number of observations in the top-right corner
+    n_obs = len(subset)
+    ax.text(
+        0.95, 0.95,
+        f"N = {n_obs}",
+        transform=ax.transAxes,
+        ha='right',
+        va='top',
+        fontsize=9,
+        color='black'
+    )
+
+# 5. Overall title
+plt.subplots_adjust(top=0.9)
+g.fig.suptitle("Yield Distribution at Woreda Level (ECDF)", fontsize=16)
+
+plt.show()
+
+# Find the percentage of values below 16 QQ/ha
+below_16 = df_avg_yield['yield'] < 16
+below_16 = df_avg_yield[(df_avg_yield['yield'] < 16) & (df_avg_yield['YEAR'] >=2015) & (df_avg_yield['YEAR'] < 2025)]
+breakpoint()
+below_16_pct = below_16.mean() * 100
+print(f"Percentage of yields below 16 QQ/ha: {below_16_pct:.2f}%")
+
+breakpoint()
 # Change W_NAME column to title case
 df_avg_yield['W_NAME'] = df_avg_yield['W_NAME'].str.title()
 

{geocif-0.1.70 → geocif-0.1.72/geocif.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: geocif
-Version: 0.1.70
+Version: 0.1.72
 Summary: Models to visualize and forecast crop conditions and yields
 Home-page: https://ritviksahajpal.github.io/yield_forecasting/
 Author: Ritvik Sahajpal
@@ -24,6 +24,10 @@ License-File: LICENSE
 [![image](https://img.shields.io/conda/vn/conda-forge/geocif.svg)](https://anaconda.org/conda-forge/geocif)
 
 
+**Generate Climatic Impact-Drivers (CIDs) from Earth Observation (EO) data**
+
+[Climatic Impact-Drivers for Crop Yield Assessment at NASA Harvest](https://www.loom.com/share/5c2dc62356c6406193cd9d9725c2a6a9)
+
 **Models to visualize and forecast crop conditions and yields**
 
 

{geocif-0.1.70 → geocif-0.1.72}/setup.py

@@ -50,6 +50,6 @@ setup(
     test_suite="tests",
     tests_require=test_requirements,
     url="https://ritviksahajpal.github.io/yield_forecasting/",
-    version="0.1.70",
+    version="0.1.72",
     zip_safe=False,
 )