From 293c1877a4652f70c55e669fb64bb7b6586983a9 Mon Sep 17 00:00:00 2001 From: nikhilmborkar Date: Wed, 26 Jul 2017 05:01:20 +0000 Subject: [PATCH] Done --- build.py | 39 ++++++++++++++++------- build.pyc | Bin 0 -> 2533 bytes tests/__init__.pyc | Bin 0 -> 174 bytes tests/test_get_categorical_variables.pyc | Bin 0 -> 2819 bytes 4 files changed, 28 insertions(+), 11 deletions(-) create mode 100644 build.pyc create mode 100644 tests/__init__.pyc create mode 100644 tests/test_get_categorical_variables.pyc diff --git a/build.py b/build.py index c7d540b..23433d1 100644 --- a/build.py +++ b/build.py @@ -1,22 +1,39 @@ -def get_categorical_variables(df): - return [] +import numpy as np +import pandas as pd +from pandas import Series, DataFrame +import operator +import matplotlib.pyplot as plt +df = pd.read_csv('data/employee_retention_data.csv') -def get_numerical_variables(df): - return [] +def get_categorical_variables(df): + return df[['dept','join_date','quit_date']] +def get_numerical_variables(df): + return df.drop(['dept','join_date','quit_date'], axis=1) def get_numerical_variables_percentile(df): - pass - + df_tp = get_numerical_variables(df) + return df_tp.describe().T def get_categorical_variables_modes(df): - pass - + categorical_df = get_categorical_variables(df) + return categorical_df.mode() def get_missing_values_count(df): - pass - + return pd.DataFrame(pd.isnull(df).sum().rename('NA_count')) def plot_histogram_with_numerical_values(df): - pass + numerical_df = get_numerical_variables(df) + plt.subplot(221) + plt.title(num_df.columns[0]) + sns.distplot(num_df.iloc[:,0], color='yellow', fit=norm, kde=False) + plt.subplot(222) + plt.title(num_df.columns[1]) + sns.distplot(num_df.iloc[:,1], color='yellow', fit=norm, kde=False) + plt.subplot(223) + plt.title(num_df.columns[2]) + sns.distplot(num_df.iloc[:,2], color='yellow', fit=norm, kde=False) + plt.subplot(224) + plt.title(num_df.columns[3]) + sns.distplot(num_df.iloc[:,3], color='yellow', fit=norm, kde=False) diff --git a/build.pyc b/build.pyc new file mode 100644 index 0000000000000000000000000000000000000000..a3a3e59187e7ba56d86ac00ce01a9ae1837ef34b GIT binary patch literal 2533 zcmcgu+m0JW5UrW9*O&EXH-G{QNT3MNylgD-93c=WkA4^me;7(L=H!6+K$?sOZz8Pr)ne16mwhTMuY6q{WDIM4K_Gh^_)!9FpqM3pkyS z>Wgwj>Ohn`qz0lKlNyS0LTV(+T@Ga)j}eyn3!hocq51=exo(VhegyTq%x909tkCXL z%oJu*UDUR^)H-dn*QGaAnezIB+&#_%81d*~e2d}kLOc`HD&%j{EGNxX(Clr~P!<8mk`Em;Mwk*V?B!wq93_$upZi&Ki>~t#+Ib$5Bvl z6X?fS2GGqeR@LK@Tm`{A=ETkt{&^CyVDyrL?5S~v5q2;oD@|3)Ea~$)Mkn?G`F%KN z$Yoog>^p!bc(o0fgXbZYIi7 z0fYy(_G2V65-%_ra8`F+4sUZge7nohO_$@_T#n!F^6;k1$!#ttx42ZecH$#sdS-rw zG#DhrkXN>9q)RTfwbc`8oR#tNyinTw#^&5z=64o+{>&dNxWnQ@h(suer?vHdfTwGh zxWv3{;ElIhHUI^tY9$8PwJl25bt#_0U&eX)6|;k}RW2vIteQgnJj$%wKI_b$uSs>^ zC;c-)SNZfDQC4eQ+36GG&v$5@+x$aD&w}agD0naU@`jAx!CQfZ@Np?}M$1|+G literal 0 HcmV?d00001 diff --git a/tests/__init__.pyc b/tests/__init__.pyc new file mode 100644 index 0000000000000000000000000000000000000000..045874b068aebf9815d3f86456a381e25fcfd731 GIT binary patch literal 174 zcmZ9GK?=e!5Je-p5W#zN6XyaVp1@5P#f>3tMjF~=LS|g_Sf0%bI0aXJ`0vACeE#-x zJiH(MeK0?5G_PvjDS|O=vo^OzVgA6Kb}&*x$Z;CuoO-I0uh405@dLs@0Re-zPpTGE pnQ1|^6$P_I$~JiMi~@oz49>+I#ko^iWUt%jMBz2#WqrVE^98D1E8zeD literal 0 HcmV?d00001 diff --git a/tests/test_get_categorical_variables.pyc b/tests/test_get_categorical_variables.pyc new file mode 100644 index 0000000000000000000000000000000000000000..2dcd747a69596600b366b5901e3e95c951be855e GIT binary patch literal 2819 zcmc&$%We}%6us?NVqgRe&q4E$Mo4&*cr`mTLIlB%MF(kwHJVnX`#N;e4^LMa$x>Ls z5Akh$0Gx9>I97-e8?ckUUAOwyh1o@iH&`l3B~i->`_TcQJVwip05z~ z3;e@%crhlXGYvhybQZ!8g`+w*&eXjB=XoH5I$}^Ds)Ko6WHFn@#1$^VIPbLyFPgcI ztd7B6i>ZYLRyFEJc!?QbqBMS0mpZQU<5zi69@NdTYsUZJt^S( zsyjvHC*1zoD_waT&l!y+)VyenL@Fo|QXp!f<6 zWs5nuX&SRx`++2FRLoy7fu)!i(C{<9dtx|~az#uo&OJsvg!v{hks-)Nto|b;BoejV z+xCsHRb{vltx+bKX_InCD%6OzOl`^LY|qFxwQ`?VKDa7X%h6dY+7qn#r;RJyBx?n+ z#v2#p0U&%$kQMaz5bC+;T5Xmo_We}(17oe-8MEqFl?_xw!U*;Yw`VdELF;}0Y^}Gt_Ipa(*_Q8iZ%L90DV!vh=n0jX7jR?l+VORo+ccyuX0n{^v{w4@XJ^c73<aS#o+ zhTFsbus>p*L7G>2AQ>?##7-2PHJS64Gj_e0V&AjSMqQThHpYDCA#N$v>dtXCj@D-R VuvO)6&d5+-i>`DBxD{OOegaqAu*v`c literal 0 HcmV?d00001