Merge pull request #369 from nasaharvest/Add-data-Senegal_2022_Set3

Senegal 2022 Set 3

data/datasets.dvc

@@ -1,6 +1,6 @@
 outs:
-- md5: 293db6eddf1a985f540e6bbc989baa81.dir
-  size: 669670393
+- md5: 63c6f22459e60941aa459671fbff891b.dir
+  size: 671350235
   nfiles: 48
   path: datasets
   hash: md5

data/raw.dvc

@@ -1,6 +1,6 @@
 outs:
-- md5: eb021dc8fddb549a048a03e29e81bfbd.dir
-  size: 443655024
-  nfiles: 386
+- md5: c033b32e0901a197d9f743c8329b24ae.dir
+  size: 443888649
+  nfiles: 387
   path: raw
   hash: md5

data/report.txt

@@ -287,14 +287,14 @@ 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%
+Senegal_CEO_2022 (Timesteps: 19)
+----------------------------------------------------------------------------
+disagreement: 0.1%
+eo_data_complete    1499
+eo_data_skipped        1
+✔ training amount: 310, positive class: 11.0%
+✔ validation amount: 579, positive class: 11.7%
+✔ testing amount: 610, positive class: 14.8%
 
 
 

datasets.py

@@ -1204,6 +1204,15 @@ def load_labels(self) -> pd.DataFrame:
                 longitude_col="lon",
                 filter_df=clean_ceo_data,
             ),
+            RawLabels(
+                filename="ceo-Senegal-March-2022---March-2023-Stratified-sample-(Set-3)-sample-data-2023-11-26.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(),