Add-data-Senegal CEO 2022 set 1&2

data/datasets.dvc

@@ -1,6 +1,6 @@
 outs:
-- md5: b6a08170b543289fc043576b00e8a65c.dir
-  size: 658002555
-  nfiles: 45
+- md5: 001feb4ecdaa108deaf43002ef840c11.dir
+  size: 663324332
+  nfiles: 46
   path: datasets
   hash: md5

data/raw.dvc

@@ -1,6 +1,6 @@
 outs:
-- md5: 53662f45a86eb8f39bd26f87f3b98e6e.dir
-  size: 442437587
-  nfiles: 380
+- md5: f63283bc4a661fb36f405f0dc99da064.dir
+  size: 442919107
+  nfiles: 382
   path: raw
   hash: md5

data/report.txt

@@ -258,6 +258,17 @@ eo_data_skipped       82
 
 
 
+Senegal_CEO_2022 (Timesteps: 16)
+----------------------------------------------------------------------------
+disagreement: 10.5%
+eo_data_complete    1342
+eo_data_skipped      158
+✔ training amount: 276, positive class: 4.7%
+✔ validation amount: 516, positive class: 6.6%
+✔ testing amount: 550, positive class: 10.7%
+
+
+
 HawaiiAgriculturalLandUse2020 (Timesteps: 24)
 ----------------------------------------------------------------------------
 eo_data_complete     4834

datasets.py

@@ -1120,6 +1120,30 @@ def load_labels(self) -> pd.DataFrame:
             ),
         ),
     ),
+    CustomLabeledDataset(
+        dataset="Senegal_CEO_2022",
+        country="Senegal",
+        raw_labels=(
+            RawLabels(
+                filename="ceo-Senegal-March-2022---March-2023-Stratified-sample-(Set-1)-sample-data-2023-08-28.csv",  # noqa: E501
+                class_prob=lambda df: (df["Does this pixel contain active cropland?"] == "Crop"),
+                start_year=2022,
+                train_val_test=(0.2, 0.4, 0.4),
+                latitude_col="lat",
+                longitude_col="lon",
+                filter_df=clean_ceo_data,
+            ),
+            RawLabels(
+                filename="ceo-Senegal-March-2022---March-2023-Stratified-sample-(Set-2)-sample-data-2023-08-28.csv",  # noqa: E501
+                class_prob=lambda df: (df["Does this pixel contain active cropland?"] == "Crop"),
+                start_year=2022,
+                train_val_test=(0.2, 0.4, 0.4),
+                latitude_col="lat",
+                longitude_col="lon",
+                filter_df=clean_ceo_data,
+            ),
+        ),
+    ),
     HawaiiAgriculturalLandUse2020(),
     KenyaCEO2019(),
     HawaiiCorrective2020(),