we are obviously using the order of sample submission only but my query is this -

Do we need to use this format(as mentioned in the discussion link mentioned above) -

pred_df = pred_df.rename(columns={

7:'Crop_ID_1',

2:'Crop_ID_2',

0:'Crop_ID_3',

1:'Crop_ID_4',

8:'Crop_ID_5',

5:'Crop_ID_6',

4:'Crop_ID_7',

6:'Crop_ID_8',

3:'Crop_ID_9'

})

or this format as mentioned in the notebook -

pred_df = pd.DataFrame(y_competition_prob) pred_df = pred_df.rename(columns={ 0:'Crop_ID_1', 1:'Crop_ID_2', 2:'Crop_ID_3', 3:'Crop_ID_4', 4:'Crop_ID_5', 5:'Crop_ID_6', 6:'Crop_ID_7', 7:'Crop_ID_8', 8:'Crop_ID_9' }) pred_df['field_id']=data_test_grouped['field_id'] pred_df = pred_df[['field_id', 'Crop_ID_1', 'Crop_ID_2', 'Crop_ID_3', 'Crop_ID_4', 'Crop_ID_5', 'Crop_ID_6', 'Crop_ID_7', 'Crop_ID_8', 'Crop_ID_9']]

Both of them are giving different answers and the scores which you see on leaderboard are mostly from following the format mentioned in the discussion link - https://zindi.africa/hackathons/radiant-earth-spot-the-crop-hackathon/discussions/6728

Yes. These two different orderings actually gave two different scores on LB respectively. One is far better than the other one. @Zindi We truly need a clarification here as regards this issue on ground.

Thanks in anticipation.

HI @ZINDI

can we please get an update on this?

Hi all,

Please follow what is on the data page

• 0 - No Data
• 1 - Lucerne/Medics
• 2 - Planted pastures (perennial)
• 3 - Fallow
• 4 - Wine grapes
• 5 - Weeds
• 6 - Small grain grazing
• 7 - Wheat
• 8 - Canola
• 9 - Rooibos

If possible @ZINDI reset the leaderboard. because most of us followed the format mentioned in the discussion link stated above and this data page format was rarely followed by anyone.

Okay. Thanks @Zindi