From 4e545821c0a34fa8d38a4e50d8d7e8b4d377f569 Mon Sep 17 00:00:00 2001
From: mudassirkhan19 <mudassirkhan19@users.noreply.github.com>
Date: Mon, 3 Jul 2017 04:13:29 +0000
Subject: [PATCH] Done

---
 build.py                                 |  45 ++++++++++++++++++-----
 build.pyc                                | Bin 0 -> 2797 bytes
 tests/__init__.pyc                       | Bin 0 -> 172 bytes
 tests/test_get_categorical_variables.pyc | Bin 0 -> 2765 bytes
 4 files changed, 35 insertions(+), 10 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 35cdd2a..f1d1944 100644
--- a/build.py
+++ b/build.py
@@ -1,26 +1,51 @@
+import matplotlib.pyplot as plt
+import seaborn as sns
+from scipy.stats import norm
+import pandas as pd
+
 def get_categorical_variables(df):
-    return []
+    df['new_user'] = df['new_user'].astype('category')
+    df['converted'] = df['converted'].astype('category')
+    cat_cols = df.select_dtypes(include=['object','category']).columns
+    return cat_cols
 
 
 def get_numerical_variables(df):
-    return []
-
+    df['new_user'] = df['new_user'].astype('category')
+    df['converted'] = df['converted'].astype('category')
+    num_cols = df.select_dtypes(exclude=['object','category']).columns
+    return num_cols
 
 def get_numerical_variables_percentile(df):
-    pass
+    df['new_user'] = df['new_user'].astype('category')
+    df['converted'] = df['converted'].astype('category')
+    num_cols = df.select_dtypes(exclude=['object','category']).columns
+    return df[num_cols].describe()
 
 
 def get_categorical_variables_modes(df):
-    pass
+    df['new_user'] = df['new_user'].astype('category')
+    df['converted'] = df['converted'].astype('category')
+    cat_cols = df.select_dtypes(include=['object','category']).columns
+    return df[cat_cols].mode()
 
 
 def get_missing_values_count(df):
-    pass
-
+    return pd.isnull(df).sum().reset_index()
 
 def plot_histogram_with_numerical_values(df):
-    pass
-
+    df['new_user'] = df['new_user'].astype('category')
+    df['converted'] = df['converted'].astype('category')
+    num_cols = df.select_dtypes(exclude=['object','category']).columns
+    plt.subplot(121)
+    plt.title(num_cols[0])
+    sns.distplot(df[num_cols[0]], color='yellow', fit=norm, kde=False)
+    plt.subplot(122)
+    plt.title(num_cols[1])
+    sns.distplot(df[num_cols[1]], color='yellow', fit=norm, kde=False)
 
 def plot_facet_box(df):
-    pass
+    plt.subplot(121)
+    sns.boxplot('converted','age',data=df)
+    plt.subplot(122)
+    sns.boxplot('converted','total_pages_visited',data=df)
diff --git a/build.pyc b/build.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..07984a59801753274ee62a2b0478264db33ceb42
GIT binary patch
literal 2797
zcmcgu+ioL85bg2AcAVII0i-2>gt$ofC2M#@S|KDPp1=={#0tNR#?u|Aot_zMx@Qw3
zcv}92PvS%P063>?E1M{Rmw25_O?CCH`gGN)N&g-l{rTn3zZ7cw>)`z@p71M#!GEH>
zQp-fmI=g17W(T5ol<%str)E8sY=`#M>`<9h{h`!spv-|NN6K_XIaa18%21iUC?jPK
zMR}ymK$OSa@~80;ws-&FHytNX;{ilfH)X*VouuJAJmE)(-^9_AC0-r%I@z{g(44B9
z=2X#)0~s8m%kc|LQOUBSe#b=a-NTC0cnDe9%d8Eykw%e6yQrIW9H6JDt5>jF6Z>en
zs_?C`ku&T0(iU-?LX~zN;(BG{2<>1!x-vsUkcma@+p-E{o+wh7v)x`9`ob6{B_jL)
zZE{|hc2c$`55YAT=Xv$j*OQlZa}ic~VJAgx?8KV<^sH`9Hy#F8SEmg;KQ(#GgN>7U
z>%4isTE`PCxUezXp34=v&tBz?%jez(o+J<gl{~}Whsj$5Uh{@EvCx7c@wHciukJz@
zWHtTM?9y~n@YB4P&9$}V-qCx8&4^yrmbSTxom};QWA_rXA1-LIR)R235y!jmy@jG4
zV#5u5Oq^Tr324TKqH%M3jiu&6B=|%;d<1@q_3yzjTiK?tRdn9Ii&6g*if<iG0a2py
zE;wl`k&ACYNj}>rlpk-Qyfc|IIuky)6UQB&3YU@+qOc=^6Fdr1L`dxdi$kYl9!{ug
zWrW{cs9NtMb2YSOJjT1RAezio#$E|~OJ?r#^zUO+22$z3i`4=&=35)U0c{mOp<|>I
zb<JUiXvu;9g7RQ_2n?1eDm@~Xu+e2+qSBp+p{JI8Znm|Zpz@*SnIm$RIaVkuuai{f
zVUgjjMMifl0^2={81ld)oNan954bhD9Uc`K$0#;+Blo$s-q)9MpJy)0`@&eq?*YdM
zkkWjcKzRErAEl;+cE0j;lx3rf-pU%G3LBbnA+lRz#%Am9b06MJm3_0$=i+M}vvaso
zFPgl}E?qpo!Q>1djtlS5&20%BB~Ozt5=tqC1cQ$R60<^3r)9c{b*g(0ecm$h%+Bxy
z5VLuGHGyRt@Z%@Y<5P$-7u3C*qx8%5u!3$>{Sim8m<F5A>!#9V6UpRIxYhc3h<OZ~
z3zJ!Sg%1!-Ue-)^O&>J3(u7O1G-`6Y{tV(e??rp~D#B!b0fuY6{f3Z-_aGi8!{kXa
P><s#Y9-jUn9UT7)X3b2q

literal 0
HcmV?d00001

diff --git a/tests/__init__.pyc b/tests/__init__.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..ca9c535ee196038abaadce79740487fa54cd922b
GIT binary patch
literal 172
zcmZ9Fu?hk)5JV$dh~R%*6BkQCM6lOFM6AM+>~Y~t_Q>uDez!kh4y+uQH!#J_{Ig!2
z+RMY6pJp|W)v~Av#$21JIX4yN4|1y=bess%7*HgZzQerT!JSJl19}3_fq(+qmW%bA
nUsNP=wgr`JKt&-1B~bzAVvg$EW_ateF7FeE@kxj2U#iU))wn7r

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..14817c875bdbff74aeb517fd5fb986d0f9c70697
GIT binary patch
literal 2765
zcmc&$O>fgc5FI-oZAwuK<y(9RghVcF#DN16LWm*|7m(aah+Hh!-n4EUJMylJkje@E
zF+Tv_8@Fj_8^i%MX?MJ{9<TSkx3lBM&&x}jgO8tMIr=qleT8QGFbV!W$wWq;^gZ$D
z`;z(6Z%ERRep5_C_FH0_@*NRvbr;06)m@YX0iFyzNm}wl<h#h{euq<Rb=#Oe{e@rG
zL(eV3Y#Zw~BWpg`JMcDoN#vqlT;vC)v}uut^fzLA;O-p-zcVh3BWHGtGL54wJc!CP
z+R2QapyAxF(d;uABLo2qSW4d~GIoH(*G`j=IMU*}*=b9#)otO)nDdYswoNo!z+eKh
zq3nDEJBRQa;(XxrzWH(}vX9A|a_GxY8S##WT4#saXNMN1h7ti~<mkf`XSd<+&Y}Rz
zzzPR$0hhR_a#wy;l5_JVtgI;=1+Ya`#>RCthCDlyber%TQSVe~mIRz9Aa1t=6InjH
zT~?;6V+^(?8)%s0P-$!z7g1w}==JuBk?D=91P`a>a4*WAKJR@h%Ap-cv4O&g>6s*2
z9~9;KiQMZYlwGIh*0`R--mq%JnUFWeU)_DYkFm3(2-Ysch_~vkcx(E1R^ayt_%8yh
zbR>IPx|{HcBd<p0)ZVSRCX*2&)^K%&!{#2twc$Z((_@hWZiwJA4Z{N>Q&j>gRo}S8
zF|;|!aTjl%$Z-zXE4Xh7<5-M|5ZqXsy!!&?pM-FlY1p~@NH!#4eO_(pexBz35hk3$
zeK<DAhuo!^QA)eao|<N^^n_;31xS%w7v;zxCdoL_M0$fA`?icmb%NAV=)-fC;t7`c
zHz}sDb9Y*9*;<tg(BuK$H8Zcn5l#quT96xf<V29*DqUtva1CZE4K!}f+&{*WCfuJ}
zv1*J=my0F-7T!qWM=8$cd>17&t8mO{#W>A4WB-M5jhi#${6v^Aejd1oe^RZPt_3m-
zb0lvVYKn){rOMz2EE&Nn&2<{)NWcnwa=L2oqAOm&PZOS?mWO`Z>nwKKomQvSrK^Cd
zJawcZz<_AkA4fTAv1SVnO_2f4I}I&Ol!W}k;wVe6CR43B-!o^dy2cl^A}_ff#k}e*
K`)yS675_J6&7Nrh

literal 0
HcmV?d00001