Merge pull request #412 from nasaharvest/Add-data-North-Uganda-2017

Created new Dataset North-Uganda-2017

data/datasets.dvc

@@ -1,6 +1,6 @@
 outs:
-- md5: b073343b2883fe4537cbdb98bd447cc2.dir
-  size: 728780256
-  nfiles: 61
+- md5: b45a782128aebf6786df801a75fbd46d.dir
+  size: 734199009
+  nfiles: 62
   path: datasets
   hash: md5

data/raw.dvc

@@ -1,6 +1,6 @@
 outs:
-- md5: 7ba2a5b698999a9bfa635252b5090ff7.dir
-  size: 446381456
-  nfiles: 406
+- md5: 0916e33f6eef6c80a87e319427005f5e.dir
+  size: 446720790
+  nfiles: 408
   path: raw
   hash: md5

data/report.txt

@@ -512,3 +512,12 @@ eo_data_export_failed    227
 ✔ training amount: 293, positive class: 16.4%
 ✔ validation amount: 235, positive class: 18.3%
 ✔ testing amount: 245, positive class: 16.7%
+
+
+
+Uganda_NorthCEO2017 (Timesteps: 24)
+----------------------------------------------------------------------------
+eo_data_complete    1000
+✔ training amount: 387, positive class: 1.3%
+✔ validation amount: 294, positive class: 1.0%
+✔ testing amount: 319, positive class: 1.3%

datasets.py

@@ -508,6 +508,38 @@ def load_labels(self) -> pd.DataFrame:
         return df
 
 
+class Uganda_NorthCEO2017(LabeledDataset):
+    def load_labels(self) -> pd.DataFrame:
+        raw_folder = raw_dir / "Uganda_North_2017"
+        df1 = pd.read_csv(
+            raw_folder
+            / "ceo-UNHCR-North-Uganda-Feb-2017---Feb-2018-(Set-1)-sample-data-2024-08-29.csv"
+        )
+        df2 = pd.read_csv(
+            raw_folder
+            / "ceo-UNHCR-North-Uganda-Feb-2017---Feb-2018-(Set-2)-sample-data-2024-08-29.csv"
+        )
+        df = pd.concat([df1, df2])
+
+        # Discard rows with no label
+        df = df[~df["Does this pixel contain active cropland?"].isna()].copy()
+        df[CLASS_PROB] = df["Does this pixel contain active cropland?"] == "Crop"
+        df[CLASS_PROB] = df[CLASS_PROB].astype(int)
+        df["num_labelers"] = 1
+        df = df.groupby([LON, LAT], as_index=False, sort=False).agg(
+            {
+                CLASS_PROB: "mean",
+                "num_labelers": "sum",
+                "plotid": join_unique,
+                "sampleid": join_unique,
+                "email": join_unique,
+            }
+        )
+        df[START], df[END] = date(2017, 1, 1), date(2018, 12, 31)
+        df[SUBSET] = train_val_test_split(df.index, 0.3, 0.3)
+        return df
+
+
 class Uganda_NorthCEO2021(LabeledDataset):
     def load_labels(self) -> pd.DataFrame:
         raw_folder = raw_dir / "Uganda_North_2021"
@@ -1538,6 +1570,7 @@ def load_labels(self) -> pd.DataFrame:
     TanzaniaCropArea2019(),
     FranceCropArea2020(),
     Uganda_NorthCEO2016(),
+    Uganda_NorthCEO2017(),
 ]
 
 if __name__ == "__main__":