In [1]:

    

from sys import path
path.append("../../")

import pandas as pd
import seaborn as sns

from matplotlib import pyplot
from utils.data_cube_utilities import dc_display_map


%matplotlib inline


    

In [2]:

    

df = pd.read_csv('../data.csv')
df.head()


    

    
Out[2]:







  

    

      

      
X
      
Y
      
LandUse
      
Latitude
      
Longitude
    
  
  

    

      
0
      
618522.8899
      
6218452.029
      
Forestry
      
-34.168155
      
-55.714061
    
    

      
1
      
649042.3861
      
6199719.402
      
Forestry
      
-34.333113
      
-55.379803
    
    

      
2
      
655993.6095
      
6200009.045
      
Forestry
      
-34.329479
      
-55.304319
    
    

      
3
      
656440.0958
      
6200236.666
      
Forestry
      
-34.327360
      
-55.299509
    
    

      
4
      
656825.7859
      
6199755.161
      
Forestry
      
-34.331642
      
-55.295231
    
  

In [3]:

    

df.groupby("LandUse").size()


    

    
Out[3]:





LandUse
Forestry              6
Fruittrees            3
Nativeforest         13
Naturalgrassland    175
Other                 8
Prairie              59
Stubble               9
Summercrops         150
Water                 1
dtype: int64

In [4]:

    

fig, ax = pyplot.subplots(figsize=(15,3))
sns.countplot(x="LandUse",data=df, palette="Greens_d");


    

In [5]:

    

df_new = df.copy()  
df_new['LandUse'].update(df_new['LandUse'].map(lambda x: "Forest" if x in ["Forestry","Fruittrees","Nativeforest"] else x ))
df_new['LandUse'].update(df_new['LandUse'].map(lambda x: "Misc" if x  not in ["Forest","Prairie","Summercrops","Naturalgrassland"] else x ))


    

In [6]:

    

df_new.groupby("LandUse").size()


    

    
Out[6]:





LandUse
Forest               22
Misc                 18
Naturalgrassland    175
Prairie              59
Summercrops         150
dtype: int64

In [7]:

    

fig, ax = pyplot.subplots(figsize=(15,5))
sns.countplot(x="LandUse",data=df_new, palette="Greens_d");


    

In [8]:

    

dc_display_map.display_grouped_pandas_rows_as_pins(df_new, group_name= "LandUse")


    

    
Out[8]:

In [9]:

    

output_destination_name = "./relabeled_data.csv"


    

In [10]:

    

## Recap of structure
df_new.head()


    

    
Out[10]:







  

    

      

      
X
      
Y
      
LandUse
      
Latitude
      
Longitude
    
  
  

    

      
0
      
618522.8899
      
6218452.029
      
Forest
      
-34.168155
      
-55.714061
    
    

      
1
      
649042.3861
      
6199719.402
      
Forest
      
-34.333113
      
-55.379803
    
    

      
2
      
655993.6095
      
6200009.045
      
Forest
      
-34.329479
      
-55.304319
    
    

      
3
      
656440.0958
      
6200236.666
      
Forest
      
-34.327360
      
-55.299509
    
    

      
4
      
656825.7859
      
6199755.161
      
Forest
      
-34.331642
      
-55.295231
    
  

In [11]:

    

df_new.to_csv(output_destination_name)


    

In [12]:

    

!ls


    

    




'1. Data Exploration.ipynb'		    __init__.py
'2. Feature Engineering.ipynb'		    classifiers
'3. Model Building and Evaluation.ipynb'    engineered_features.csv
'4. Display and Package Classifier.ipynb'   engineered_features_original.csv
'Old Notebooks'				    feature_engineering.PNG
 README.md				    relabeled_data.csv