Develop Reference

Creating a categorical variable from two dummy variables

I have the following data;
{'ID': {0: 5531.0, 1: 2658.0, 2: 5365.0, 3: 4468.0, 4: 3142.0},
'FEMALE': {0: 1.0, 1: 0.0, 2: 0.0, 3: 0.0, 4: 1.0},
'MALE': {0: 0.0, 1: 1.0, 2: 1.0, 3: 1.0, 4: 0.0},
'AGE': {0: 45.0, 1: 40.0, 2: 38.0, 3: 43.0, 4: 38.0},
'S': {0: 12.0, 1: 12.0, 2: 15.0, 3: 13.0, 4: 18.0}}
Where MALE is a dummy equal to one if the individual is male, 0 otherwise. The same for FEMALE.
I want to create a new variable, Gender, which is categorical. If MALE==1 then Gender = Male, if FEMALE==1 then Gender = Female. The purpose is to allow for a clear twoway scatter plot seperated by gender. I can do this currently, but the legend is hard to understand.
I tried the following;
import numpy as np
import pandas as pd
stata_data_P1 = pd.DataFrame({'ID': {0: 5531.0, 1: 2658.0, 2: 5365.0, 3: 4468.0, 4: 3142.0}, 'FEMALE': {0: 1.0, 1: 0.0, 2: 0.0, 3: 0.0, 4: 1.0}, 'MALE': {0: 0.0, 1: 1.0, 2: 1.0, 3: 1.0, 4: 0.0}, 'AGE': {0: 45.0, 1: 40.0, 2: 38.0, 3: 43.0, 4: 38.0}, 'S': {0: 12.0, 1: 12.0, 2: 15.0, 3: 13.0, 4: 18.0}})
stata_data_P1['Gender'] = np.where(stata_data_P1['MALE'] == '1', 'Female', 'Male')
stata_data_P1.head()
But from stata_data_P1.head() we can see it doesn't seem to have taken on board my command for true and false values.
Any Help would be greatly appreciated.

First use assign method to create new column then use idxmax just in MALE and FEMALE columns to return the index for the maximum value in each row.
Code:
stata_data_P1.assign(GENDER=lambda df_: df_.loc[:, ["MALE", "FEMALE"]].idxmax(axis=1))
Documentation:
Pandas - idxmax
Pandas - assign

Pandas : Fillna for all columns, except two

I wonder how can we fill the NaNs from all columns of a dataframe, except some.
For example, I have a dataframe with 20 columns, I want to fill the NaN for all except two columns (in my case, NaN are replaced by the mean).
df = df.drop(['col1','col2], 1).fillna(df.mean())
I tried this, but I don't think it's the best way to achieve this (also, i want to avoid the inplace=true arg).
Thank's

You can select which columns to use fillna on. Assuming you have 20 columns and you want to fill all of them except 'col1' and 'col2' you can create a list with the ones you want to fill:
f = [c for c in df.columns if c not in ['col1','col2']]
df[f] = df[f].fillna(df[f].mean())
print(df)
col1 col2 col3 col4 ... col17 col18 col19 col20
0 1.0 1.0 1.000000 1.0 ... 1.000000 1 1.000000 1
1 NaN NaN 2.666667 2.0 ... 2.000000 2 2.000000 2
2 NaN 3.0 3.000000 1.5 ... 2.333333 3 2.333333 3
3 4.0 4.0 4.000000 1.5 ... 4.000000 4 4.000000 4
(2.66666) was the mean
# Initial DF:
{'col1': {0: 1.0, 1: nan, 2: nan, 3: 4.0},
'col2': {0: 1.0, 1: nan, 2: 3.0, 3: 4.0},
'col3': {0: 1.0, 1: nan, 2: 3.0, 3: 4.0},
'col4': {0: 1.0, 1: 2.0, 2: nan, 3: nan},
'col5': {0: 1.0, 1: nan, 2: 3.0, 3: 4.0},
'col6': {0: 1, 1: 2, 2: 3, 3: 4},
'col7': {0: nan, 1: 2.0, 2: 3.0, 3: 4.0},
'col8': {0: 1.0, 1: 2.0, 2: nan, 3: 4.0},
'col9': {0: 1, 1: 2, 2: 3, 3: 4},
'col10': {0: 1.0, 1: 2.0, 2: nan, 3: 4.0},
'col11': {0: 1.0, 1: nan, 2: 3.0, 3: 4.0},
'col12': {0: 1, 1: 2, 2: 3, 3: 4},
'col13': {0: 1.0, 1: 2.0, 2: nan, 3: 4.0},
'col14': {0: 1.0, 1: nan, 2: 3.0, 3: 4.0},
'col15': {0: 1, 1: 2, 2: 3, 3: 4},
'col16': {0: 1.0, 1: nan, 2: 3.0, 3: nan},
'col17': {0: 1.0, 1: 2.0, 2: nan, 3: 4.0},
'col18': {0: 1, 1: 2, 2: 3, 3: 4},
'col19': {0: 1.0, 1: 2.0, 2: nan, 3: 4.0},
'col20': {0: 1, 1: 2, 2: 3, 3: 4}}

Reshape error when using mutual_info regression for feature selection

I am trying to do some feature selection using mutual_info_regression with SelectKBest wrapper. However I keep running into an error indicating that my list of features needs to be reshaped into a 2D array, not quite sure why I keep getting this message-
#feature selection before linear regression benchmark test
import sklearn
from sklearn.feature_selection import mutual_info_regression, SelectKBest
features = list(housing_data[housing_data.columns.difference(['sale_price'])])
target = 'sale_price'
new = SelectKBest(mutual_info_regression, k=20).fit_transform(features, target)
This is my traceback:
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
<ipython-input-18-8c778124066c> in <module>()
3 features = list(housing_data[housing_data.columns.difference(['sale_price'])])
4 target = 'sale_price'
----> 5 new = SelectKBest(mutual_info_regression, k=20).fit_transform(features, target)
/usr/local/lib/python3.6/dist-packages/sklearn/base.py in fit_transform(self, X, y, **fit_params)
463 else:
464 # fit method of arity 2 (supervised transformation)
--> 465 return self.fit(X, y, **fit_params).transform(X)
466
467
/usr/local/lib/python3.6/dist-packages/sklearn/feature_selection/univariate_selection.py in fit(self, X, y)
339 self : object
340 """
--> 341 X, y = check_X_y(X, y, ['csr', 'csc'], multi_output=True)
342
343 if not callable(self.score_func):
/usr/local/lib/python3.6/dist-packages/sklearn/utils/validation.py in check_X_y(X, y, accept_sparse, accept_large_sparse, dtype, order, copy, force_all_finite, ensure_2d, allow_nd, multi_output, ensure_min_samples, ensure_min_features, y_numeric, warn_on_dtype, estimator)
754 ensure_min_features=ensure_min_features,
755 warn_on_dtype=warn_on_dtype,
--> 756 estimator=estimator)
757 if multi_output:
758 y = check_array(y, 'csr', force_all_finite=True, ensure_2d=False,
/usr/local/lib/python3.6/dist-packages/sklearn/utils/validation.py in check_array(array, accept_sparse, accept_large_sparse, dtype, order, copy, force_all_finite, ensure_2d, allow_nd, ensure_min_samples, ensure_min_features, warn_on_dtype, estimator)
550 "Reshape your data either using array.reshape(-1, 1) if "
551 "your data has a single feature or array.reshape(1, -1) "
--> 552 "if it contains a single sample.".format(array))
553
554 # in the future np.flexible dtypes will be handled like object dtypes
ValueError: Expected 2D array, got 1D array instead:
array=['APPBBL' 'APPDate' 'Address' 'AreaSource' 'AssessLand' 'AssessTot' 'BBL'
'BldgArea' 'BldgClass' 'BldgDepth' 'BldgFront' 'BoroCode' 'Borough'
'BsmtCode' 'BuiltFAR' 'CB2010' 'CD' 'CT2010' 'ComArea' 'CommFAR'
'CondoNo' 'Council' 'EDesigNum' 'Easements' 'ExemptLand' 'ExemptTot'
'Ext' 'FIRM07_FLA' 'FacilFAR' 'FactryArea' 'FireComp' 'GarageArea'
'HealthArea' 'HealthCent' 'HistDist' 'IrrLotCode' 'LandUse' 'Landmark'
'LotArea' 'LotDepth' 'LotFront' 'LotType' 'LtdHeight' 'MAPPLUTO_F'
'NumBldgs' 'NumFloors' 'OfficeArea' 'OtherArea' 'Overlay1' 'Overlay2'
'OwnerName' 'OwnerType' 'PFIRM15_FL' 'PLUTOMapID' 'PolicePrct' 'ProxCode'
'ResArea' 'ResidFAR' 'RetailArea' 'SHAPE_Area' 'SHAPE_Leng' 'SPDist1'
'SPDist2' 'SPDist3' 'Sanborn' 'SanitBoro' 'SanitDistr' 'SanitSub'
'SchoolDist' 'SplitZone' 'StrgeArea' 'TaxMap' 'Tract2010' 'UnitsRes'
'UnitsTotal' 'Unnamed: 0' 'Version' 'XCoord' 'YCoord' 'YearAlter1'
'YearAlter2' 'YearBuilt' 'ZMCode' 'ZipCode' 'ZoneDist1' 'ZoneDist2'
'ZoneDist3' 'ZoneDist4' 'ZoneMap' 'address' 'apartment_number' 'block'
'borough' 'building_class' 'building_class_at_sale'
'building_class_category' 'commercial_units' 'easement' 'gross_sqft'
'land_sqft' 'lot' 'neighborhood' 'price_range' 'residential_units'
'sale_date' 'tax_class' 'tax_class_at_sale' 'total_units' 'year_built'
'year_of_sale' 'zip_code'].
Reshape your data either using array.reshape(-1, 1) if your data has a single feature or array.reshape(1, -1) if it contains a single sample.
Here is a sample of my data:
housing_data = pd.DataFrame({'Unnamed: 0': {0: 1, 1: 2, 2: 3, 3: 4}, 'borough': {0: 3, 1: 3, 2: 3, 3: 3}, 'neighborhood': {0: 'DOWNTOWN-METROTECH', 1: 'DOWNTOWN-FULTON FERRY', 2: 'BROOKLYN HEIGHTS', 3: 'MILL BASIN'}, 'building_class_category': {0: '28 COMMERCIAL CONDOS', 1: '29 COMMERCIAL GARAGES', 2: '21 OFFICE BUILDINGS', 3: '22 STORE BUILDINGS'}, 'tax_class': {0: '4', 1: '4', 2: '4', 3: '4'}, 'block': {0: 140, 1: 54, 2: 204, 3: 8470}, 'lot': {0: 1001, 1: 1, 2: 1, 3: 55}, 'easement': {0: nan, 1: nan, 2: nan, 3: nan}, 'building_class': {0: 'R5', 1: 'G7', 2: 'O6', 3: 'K6'}, 'address': {0: '330 JAY STREET', 1: '85 JAY STREET', 2: '29 COLUMBIA HEIGHTS', 3: '5120 AVENUE U'}, 'apartment_number': {0: 'COURT', 1: nan, 2: nan, 3: nan}, 'zip_code': {0: 11201, 1: 11201, 2: 11201, 3: 11234}, 'residential_units': {0: 0, 1: 0, 2: 0, 3: 0}, 'commercial_units': {0: 1, 1: 0, 2: 0, 3: 123}, 'total_units': {0: 1, 1: 0, 2: 0, 3: 123}, 'land_sqft': {0: 0.0, 1: 134988.0, 2: 32000.0, 3: 905000.0}, 'gross_sqft': {0: 0.0, 1: 0.0, 2: 304650.0, 3: 2548000.0}, 'year_built': {0: 2002, 1: 0, 2: 1924, 3: 1970}, 'tax_class_at_sale': {0: 4, 1: 4, 2: 4, 3: 4}, 'building_class_at_sale': {0: 'R5', 1: 'G7', 2: 'O6', 3: 'K6'}, 'sale_price': {0: 499401179.0, 1: 345000000.0, 2: 340000000.0, 3: 276947000.0}, 'sale_date': {0: '2008-04-23', 1: '2016-12-20', 2: '2016-08-03', 3: '2012-11-28'}, 'year_of_sale': {0: 2008, 1: 2016, 2: 2016, 3: 2012}, 'Borough': {0: nan, 1: 'BK', 2: 'BK', 3: 'BK'}, 'CD': {0: nan, 1: 302.0, 2: 302.0, 3: 318.0}, 'CT2010': {0: nan, 1: 21.0, 2: 1.0, 3: 698.0}, 'CB2010': {0: nan, 1: 3017.0, 2: 1003.0, 3: 2005.0}, 'SchoolDist': {0: nan, 1: 13.0, 2: 13.0, 3: 22.0}, 'Council': {0: nan, 1: 33.0, 2: 33.0, 3: 46.0}, 'ZipCode': {0: nan, 1: 11201.0, 2: 11201.0, 3: 11234.0}, 'FireComp': {0: nan, 1: 'L118', 2: 'E205', 3: 'E323'}, 'PolicePrct': {0: nan, 1: 84.0, 2: 84.0, 3: 63.0}, 'HealthCent': {0: nan, 1: 36.0, 2: 38.0, 3: 35.0}, 'HealthArea': {0: nan, 1: 1000.0, 2: 2300.0, 3: 8822.0}, 'SanitBoro': {0: nan, 1: 3.0, 2: 3.0, 3: 3.0}, 'SanitDistr': {0: nan, 1: 2.0, 2: 2.0, 3: 18.0}, 'SanitSub': {0: nan, 1: '1B', 2: '1A', 3: '4E'}, 'Address': {0: nan, 1: '87 JAY STREET', 2: '29 COLUMBIA HEIGHTS', 3: '5120 AVENUE U'}, 'ZoneDist1': {0: nan, 1: 'M1-2/R8', 2: 'M2-1', 3: 'M3-1'}, 'ZoneDist2': {0: nan, 1: nan, 2: nan, 3: nan}, 'ZoneDist3': {0: nan, 1: nan, 2: nan, 3: nan}, 'ZoneDist4': {0: nan, 1: nan, 2: nan, 3: nan}, 'Overlay1': {0: nan, 1: nan, 2: nan, 3: nan}, 'Overlay2': {0: nan, 1: nan, 2: nan, 3: nan}, 'SPDist1': {0: nan, 1: 'MX-2', 2: nan, 3: nan}, 'SPDist2': {0: nan, 1: nan, 2: nan, 3: nan}, 'SPDist3': {0: nan, 1: nan, 2: nan, 3: nan}, 'LtdHeight': {0: nan, 1: nan, 2: nan, 3: nan}, 'SplitZone': {0: nan, 1: 'N', 2: 'N', 3: 'N'}, 'BldgClass': {0: nan, 1: 'G7', 2: 'O6', 3: 'K6'}, 'LandUse': {0: nan, 1: 10.0, 2: 5.0, 3: 5.0}, 'Easements': {0: nan, 1: 0.0, 2: 0.0, 3: 1.0}, 'OwnerType': {0: nan, 1: 'P', 2: nan, 3: nan}, 'OwnerName': {0: nan, 1: '85 JAY STREET BROOKLY', 2: '25-30 COLUMBIA HEIGHT', 3: 'BROOKLYN KINGS PLAZA'}, 'LotArea': {0: nan, 1: 134988.0, 2: 32000.0, 3: 905000.0}, 'BldgArea': {0: nan, 1: 0.0, 2: 304650.0, 3: 2548000.0}, 'ComArea': {0: nan, 1: 0.0, 2: 304650.0, 3: 2548000.0}, 'ResArea': {0: nan, 1: 0.0, 2: 0.0, 3: 0.0}, 'OfficeArea': {0: nan, 1: 0.0, 2: 264750.0, 3: 0.0}, 'RetailArea': {0: nan, 1: 0.0, 2: 0.0, 3: 1263000.0}, 'GarageArea': {0: nan, 1: 0.0, 2: 0.0, 3: 1285000.0}, 'StrgeArea': {0: nan, 1: 0.0, 2: 0.0, 3: 0.0}, 'FactryArea': {0: nan, 1: 0.0, 2: 0.0, 3: 0.0}, 'OtherArea': {0: nan, 1: 0.0, 2: 39900.0, 3: 0.0}, 'AreaSource': {0: nan, 1: 7.0, 2: 2.0, 3: 2.0}, 'NumBldgs': {0: nan, 1: 0.0, 2: 1.0, 3: 4.0}, 'NumFloors': {0: nan, 1: 0.0, 2: 13.0, 3: 2.0}, 'UnitsRes': {0: nan, 1: 0.0, 2: 0.0, 3: 0.0}, 'UnitsTotal': {0: nan, 1: 0.0, 2: 0.0, 3: 123.0}, 'LotFront': {0: nan, 1: 490.5, 2: 92.42, 3: 930.0}, 'LotDepth': {0: nan, 1: 275.33, 2: 335.92, 3: 859.0}, 'BldgFront': {0: nan, 1: 0.0, 2: 335.0, 3: 0.0}, 'BldgDepth': {0: nan, 1: 0.0, 2: 92.0, 3: 0.0}, 'Ext': {0: nan, 1: nan, 2: nan, 3: nan}, 'ProxCode': {0: nan, 1: 0.0, 2: 0.0, 3: 0.0}, 'IrrLotCode': {0: nan, 1: 'N', 2: 'Y', 3: 'Y'}, 'LotType': {0: nan, 1: 5.0, 2: 3.0, 3: 3.0}, 'BsmtCode': {0: nan, 1: 5.0, 2: 5.0, 3: 5.0}, 'AssessLand': {0: nan, 1: 1571850.0, 2: 1548000.0, 3: 36532350.0}, 'AssessTot': {0: nan, 1: 1571850.0, 2: 25463250.0, 3: 149792400.0}, 'ExemptLand': {0: nan, 1: 1571850.0, 2: 0.0, 3: 0.0}, 'ExemptTot': {0: nan, 1: 1571850.0, 2: 0.0, 3: 0.0}, 'YearBuilt': {0: nan, 1: 0.0, 2: 1924.0, 3: 1970.0}, 'YearAlter1': {0: nan, 1: 0.0, 2: 1980.0, 3: 0.0}, 'YearAlter2': {0: nan, 1: 0.0, 2: 0.0, 3: 0.0}, 'HistDist': {0: nan, 1: nan, 2: nan, 3: nan}, 'Landmark': {0: nan, 1: nan, 2: nan, 3: nan}, 'BuiltFAR': {0: nan, 1: 0.0, 2: 9.52, 3: 2.82}, 'ResidFAR': {0: nan, 1: 7.2, 2: 0.0, 3: 0.0}, 'CommFAR': {0: nan, 1: 2.0, 2: 2.0, 3: 2.0}, 'FacilFAR': {0: nan, 1: 6.5, 2: 0.0, 3: 0.0}, 'BoroCode': {0: nan, 1: 3.0, 2: 3.0, 3: 3.0}, 'BBL': {0: nan, 1: 3000540001.0, 2: 3002040001.0, 3: 3084700055.0}, 'CondoNo': {0: nan, 1: 0.0, 2: 0.0, 3: 0.0}, 'Tract2010': {0: nan, 1: 21.0, 2: 1.0, 3: 698.0}, 'XCoord': {0: nan, 1: 988208.0, 2: 985952.0, 3: 1006597.0}, 'YCoord': {0: nan, 1: 195011.0, 2: 195007.0, 3: 161424.0}, 'ZoneMap': {0: nan, 1: '12d', 2: '12d', 3: '23b'}, 'ZMCode': {0: nan, 1: nan, 2: nan, 3: nan}, 'Sanborn': {0: nan, 1: '302 016', 2: '302 004', 3: '319 077'}, 'TaxMap': {0: nan, 1: 30101.0, 2: 30106.0, 3: 32502.0}, 'EDesigNum': {0: nan, 1: nan, 2: nan, 3: nan}, 'APPBBL': {0: nan, 1: 3000540001.0, 2: 0.0, 3: 0.0}, 'APPDate': {0: nan, 1: '12/06/2002', 2: nan, 3: nan}, 'PLUTOMapID': {0: nan, 1: 1.0, 2: 1.0, 3: 1.0}, 'FIRM07_FLA': {0: nan, 1: nan, 2: nan, 3: 1.0}, 'PFIRM15_FL': {0: nan, 1: nan, 2: nan, 3: 1.0}, 'Version': {0: nan, 1: '17V1.1', 2: '17V1.1', 3: '17V1.1'}, 'MAPPLUTO_F': {0: nan, 1: 0.0, 2: 0.0, 3: 0.0}, 'SHAPE_Leng': {0: nan, 1: 1559.88914353, 2: 890.718521021, 3: 3729.78685686}, 'SHAPE_Area': {0: nan, 1: 140131.577176, 2: 34656.4472405, 3: 797554.847834}, 'price_range': {0: nan, 1: nan, 2: nan, 3: nan}})

The transformer expects a 2D array, of shape (n x m) where n is the number of samples and m the number of features and if you look at the shape of features I imagine it will display: (m,).
Reshaping arrays
In general for a feature array of shape (n,), you can do as the error code suggests and call .reshape(-1,1) on your feature array, the -1 lets it infer the additional dimension: The shape of the array will be (n,m), where for a 1 feature case m = 1.
Sklearn transformers
The above being said, I think there is additional errors with your code and understanding.
I would print features to screen and check it is what you want, it looks like you are printing a list of all the column names except sale_price.
I am not familiar with SelectKBest but it requires an (n,m) feature array not a list of column names of the features.
Additionally, target should not be the name of the target column, but an array of shape (n,), where its values are the observed target values of the training instances.
I would suggest checking the documentation (previously referenced) while you are writing your code to make sure you are using the correct arguments and utilising the function as it is intended.
Extracting features
Your data seems in a strange format (dictionary's nested in a pandas DF). However is a explicit example of how I would extract features from a pd.DataFrame for use with functions from the SKlearn framework.
housing_data = pd.DataFrame({'age': [1,5,1,10], 'size':[0,1,2,0],
'price':[190,100,50,100]
})
feature_arr = housing_data.drop('price', axis=1).values
target_values = housing_data['price']
Print feature_arr and you will hopefully see your issue. Normally you would then have to preprocess the data to, for example, drop NaN values or perform feature scaling.

'forward replace' with pandas

I have a df where every row looks something like this:
series = pd.Series({0: 1.0,
1: 1.0,
10: nan,
11: nan,
2: 1.0,
3: 1.0,
4: nan,
5: nan,
6: nan,
7: nan,
8: nan,
9: nan,
'B': 3.0,
'D': 1.0})
For every value in D I want series[series.D] to be 2 and, all numerical values that are higher than series.D, also to be 2. It's kinda like a 'forward replace'.
So what I want is:
target = pd.Series({0: 1.0,
1: 2.0,
10: nan,
11: nan,
2: 2.0,
3: 2.0,
4: nan,
5: nan,
6: nan,
7: nan,
8: nan,
9: nan,
'B': 3.0,
'D': 1.0})
So far I've got:
def forward_replace(series):
if pd.notnull(series['D']):
cols = [0,1,2,3,4,5,6,7,8,9,10,11]
target_cols = [x for x in cols if x > series.D]
series.loc[target_cols].replace({1:2}, inplace=True)
return series
It seems like it's not possible to use label based indexing with numerical column labels?

Applying a lambda function with three arguments within a Group By

Currently attempting to create a function where I divide columns in my DataFrame called DF_1 and group them by a dimension column in the same DataFrame.
The below code is attempting to achieve this by first grouping by the dimension column and applying a lambda function to each of the columns that I am trying to divide in order to get the average of each of the metrics i.e. cost per conversions, or cost per click.
Unfortunately, I am unsure how to accomplish this. The below code gives an error of TypeError: lambda() takes 2 positional arguments but 3 were given
calc_1 = DF_1[['Conversions_10D', 'Total_Revenue', 'Total_Revenue', 'Clicks', 'Spend']]
calc_2 = DF_1[['Impressions', 'Spend', 'Conversions_10D', 'Impressions', 'Clicks' ]]
def agg_avg(df, group_field, list_a, list_b):
grouped = df.groupby(group_field, as_index = False).apply(lambda x, y: x/y, list_a, list_b)
grouped = pd.DataFrame(grouped).reset_index(drop = True)
return grouped
{'Date': {0: '2018-02-28', 1: '2018-02-28', 2: '2018-02-28', 3: '2018-02-28', 4: '2018-02-28'}, 'Audience_Category': {0: 'Affinity', 1: 'Affinity', 2: 'Affinity', 3: 'Affinity', 4: 'Affinity'},
'Demo': {0: 'F25-34', 1: 'F25-34', 2: 'F25-34', 3: 'F25-34', 4: 'F25-34'}, 'Gender': {0: 'Female', 1: 'Female', 2: 'Female', 3: 'Female', 4: 'Female'},
'Device': {0: 'Android', 1: 'Android', 2: 'Android', 3: 'Android', 4: 'Android'},
'Creative': {0: 'Bubble:15', 1: 'Bubble:30', 2: 'Wide :15', 3: 'Oscar :15', 4: 'Oscar :30'},
'Impressions': {0: 3834, 1: 3588, 2: 3831, 3: 3876, 4: 3676},
'Clicks': {0: 2.0, 1: 0.0, 2: 4.0, 3: 2.0, 4: 1.0},
'Conversions_10D': {0: 0.0, 1: 0.0, 2: 0.0, 3: 0.0, 4: 0.0},
'Total_Revenue': {0: 0.0, 1: 0.0, 2: 0.0, 3: 0.0, 4: 0.0}, 'Spend': {0: 28.600707059999991, 1: 25.95319236000001, 2: 28.29383795999998, 3: 29.287063200000013, 4: 26.514734159999968},
'Demo_Category': {0: 'Narrow', 1: 'Broad', 2: 'Narrow', 3: 'Broad', 4: 'Narrow'}
'CPM_Efficiency': {0: 'Low CPM', 1: 'Low CPM', 2: 'Low CPM', 3: 'Low CPM', 4: 'Low CPM'}}