From 307caf38738bf50e3b9cd55dd64f02ebfe3e325b Mon Sep 17 00:00:00 2001
From: tracedence <shoaibrashid887@gmail.com>
Date: Thu, 24 Jan 2019 14:24:07 +0000
Subject: [PATCH 1/4] Done

---
 __pycache__/__init__.cpython-36.pyc           | Bin 159 -> 154 bytes
 .../__pycache__/__init__.cpython-36.pyc       | Bin 170 -> 174 bytes
 .../__pycache__/build.cpython-36.pyc          | Bin 655 -> 573 bytes
 q01_outlier_removal/build.py                  |  17 +++++++++++++++++
 .../tests/__pycache__/__init__.cpython-36.pyc | Bin 185 -> 180 bytes
 .../test_q01_outlier_removal.cpython-36.pyc   | Bin 1835 -> 1834 bytes
 6 files changed, 17 insertions(+)

diff --git a/__pycache__/__init__.cpython-36.pyc b/__pycache__/__init__.cpython-36.pyc
index 2e5da7da778c48a9c58d7d055bbdb642fcb66418..5386a134d1382a2cbfcabb9657ee61757955bfed 100644
GIT binary patch
delta 52
zcmbQwIE#_Pn3tDJcCTm5L=JNWcm0g~+*JLNqQvCXl+?WBRQ>S$qU_=VAXh&*KP7cy
GiZuY6TM<40

delta 57
zcmbQmIG>Tjn3tDpxuj9lL=JORAN|nc)S_bj#H5VO;*8Yn;?ks|#N5QZ%)~tXywvje
Mw9K5;;)(Ir0Qch)egFUf

diff --git a/q01_outlier_removal/__pycache__/__init__.cpython-36.pyc b/q01_outlier_removal/__pycache__/__init__.cpython-36.pyc
index 2f9a42a105b2b26ec10c60ac4d11fa03f7624d22..3dca5359b0e3664e6eb87560ae6ff2332ea56de2 100644
GIT binary patch
delta 52
zcmZ3*xQ>y-n3tDJcCTm5L=JO>X#I@*+*JLNqQvCXl+?WBRQ>S$qU_=VAXh&*KP7cy
Gv_Al(JrR5W

delta 48
zcmZ3-xQda(n3tF9Mtoz`L=JPAaQ%$@+*JL<oYV?^m(=3ylKcYwfTH}Y)Z~)liShmb
Da?lWg

diff --git a/q01_outlier_removal/__pycache__/build.cpython-36.pyc b/q01_outlier_removal/__pycache__/build.cpython-36.pyc
index 8248a1626e901c868c177407620fc7ec379db180..16d8794695f27a844b861419bcec096a06376abf 100644
GIT binary patch
delta 365
zcmeBY-OIvd%*)GFbjmZPm1QDVXFU@G0|N^XI|FfX7?7x8sAa4H!W2eH1_+tOkiyi=
zn9U#wlVJwS07;+>3yKUYSO!R<>Ir7hWb=#ScPuE#$xKenEAh-r&d*Iv=eoYQzQC62
zy?vCRbABS4WK92|uNpBxIWC|4#5~8`{L;J<O~zXsg{43PGILTjnQpP9B$g!J;>-cc
z$Ag$F8H(6}KKK=(p8+&pzoaNJIW;9UFF92|JijQrxB$r2PtH$C)z8UK&nzy<OpY&_
ze2>vylndl{5W&F2!YBoTOhufNEtq6%*nna{Yl`v<ia3Exu<=L)2S}R3CO1E&G$+-L
Q8OSRJ>EK}EV&q`~0Nw&#ZvX%Q

literal 655
zcmZ`%F>ll`6t?3ecRlsEGM>c5b_@-O6#-SdR6-RJOj#l~&m2yDwre|8i^6pL1ZH;r
z50h7>F8u*aJcofR5_aCFJb&+dFYh_e=kwVQ`K`Xo2>D5N7Q^?~X#NQT5TJ_GRM0Mi
zY)fk<SeFBXe4Et_3Mjvjx&Zo)tR}Y@PtHC)1|E2&4A+fy0%RMc(M_*gCAb!F*!rtD
zd6VeVlUs_CHbGiBT7h<q=D#8gSba^Gq$8>JXNo@C@iX-E9iIg{&><VxmQmDp$fpcT
zGEgX%<RZsR4D<Zu7;@QvIE)^!!}~hyvplO8N7gFYave@}Yc^v2VO)2w>W9aM-{;~i
z^?I~1y$&vQy2|79vPZ2{A{N$39byj7*jPvn;u`tAae8Tvk>Yb6Y5SeHIf{S1uRdbI
ziYswlod|ysjIG`}(}^~C9Qs=NAlt@?wG-Z_!P`G%b@}vJWBQ<^a7b^=6<5`HFBKfx
zo4ELg*-JNYfI#TMxcBL!ak~Yct;$$hu7UfQS%_07I5c>zF3H>^K?)l`cn2QfS)9Ju
T80eLFiN783CBihD(f#~4xYMeW

diff --git a/q01_outlier_removal/build.py b/q01_outlier_removal/build.py
index ec278ba..41105c5 100644
--- a/q01_outlier_removal/build.py
+++ b/q01_outlier_removal/build.py
@@ -1,3 +1,4 @@
+# %load q01_outlier_removal/build.py
 # Default imports
 import pandas as pd
 
@@ -6,3 +7,19 @@
 
 
 # Write your Solution here:
+def outlier_removal(data):
+    
+    loan_data = data
+    loan_data = loan_data[loan_data['ApplicantIncome'] < loan_data['ApplicantIncome'].quantile(0.97)]
+    loan_data = loan_data[loan_data['CoapplicantIncome'] < loan_data['CoapplicantIncome'].quantile(0.98)]
+    loan_data = loan_data[loan_data['LoanAmount'] < loan_data['LoanAmount'].quantile(0.97)]
+    
+    return loan_data
+# loan_data = loan_data[loan_data['ApplicantIncome'] < loan_data['ApplicantIncome'].quantile(0.97)]
+# loan_data = loan_data[loan_data['CoapplicantIncome'] < loan_data['CoapplicantIncome'].quantile(0.98)]
+# loan_data = loan_data[loan_data['LoanAmount'] < loan_data['LoanAmount'].quantile(0.97)]
+# loan_data = loan_data[(loan_data['ApplicantIncome'] < loan_data['ApplicantIncome'].quantile(0.95)) & (loan_data['CoapplicantIncome'] < loan_data['CoapplicantIncome'].quantile(0.95)) & (loan_data['LoanAmount'] < loan_data['LoanAmount'].quantile(0.95))]
+# loan_data.shape
+
+
+
diff --git a/q01_outlier_removal/tests/__pycache__/__init__.cpython-36.pyc b/q01_outlier_removal/tests/__pycache__/__init__.cpython-36.pyc
index 5a057ffb73694628cef3ed87e03ee3a17f7410bc..4ddb6092a9acfd5ca51e4af57dd75244637d19c4 100644
GIT binary patch
delta 52
zcmdnVxP_6!n3tDJcCTm5L=JO>ME#8X+*JLNqQvCXl+?WBRQ>S$qU_=VAXh&*KP7cy
GN+<xSUlFDN

delta 57
zcmdnOxRa5?n3tDpxuj9lL=JP+H2u)x)S_bj#H5VO;*8Yn;?ks|#N5QZ%)~tXywvje
Mw9K5;;)(I000(>%<^TWy

diff --git a/q01_outlier_removal/tests/__pycache__/test_q01_outlier_removal.cpython-36.pyc b/q01_outlier_removal/tests/__pycache__/test_q01_outlier_removal.cpython-36.pyc
index 4c0b6c7431c6a14108ba9d55dd44059612f144a5..c92128b985ab5e21687eb9472b271efc6efdc784 100644
GIT binary patch
delta 68
zcmZ3@w~CL`n3tDJcCTm5$%&k+IQY_2OVUbnayD*vVN@v8&&bbB)h{VZOioQn%}Y+z
X56>^kE-nCa^^@~cQa39wakByd8xa>r

delta 69
zcmZ3*x0;XBn3tE!Sjs4B??ld3EL`cSB^!6UFshd6hZd(673(J^Wn>m-q-GbFCKV;-
ZCgx=(=IQ69mdB@M=A;&HmS^H-1po|o7gYcN


From c4f1e5195c30fc56effaaaead3704c408780dcc7 Mon Sep 17 00:00:00 2001
From: tracedence <shoaibrashid887@gmail.com>
Date: Fri, 25 Jan 2019 14:08:57 +0000
Subject: [PATCH 2/4] Done

---
 .../__pycache__/__init__.cpython-36.pyc        | Bin 0 -> 176 bytes
 .../__pycache__/build.cpython-36.pyc           | Bin 0 -> 1133 bytes
 q02_data_cleaning_all/build.py                 |  17 +++++++++++++++++
 .../tests/__pycache__/__init__.cpython-36.pyc  | Bin 0 -> 182 bytes
 .../test_q02_data_cleaning.cpython-36.pyc      | Bin 0 -> 3404 bytes
 5 files changed, 17 insertions(+)
 create mode 100644 q02_data_cleaning_all/__pycache__/__init__.cpython-36.pyc
 create mode 100644 q02_data_cleaning_all/__pycache__/build.cpython-36.pyc
 create mode 100644 q02_data_cleaning_all/tests/__pycache__/__init__.cpython-36.pyc
 create mode 100644 q02_data_cleaning_all/tests/__pycache__/test_q02_data_cleaning.cpython-36.pyc

diff --git a/q02_data_cleaning_all/__pycache__/__init__.cpython-36.pyc b/q02_data_cleaning_all/__pycache__/__init__.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..0b6a2eb2230e6a13a8c077608723c19c05ce7276
GIT binary patch
literal 176
zcmXr!<>ivy>lwoU1dl-k3@`#24nSPY0whuxf*CX!{Z=v*frJsnuUP$z{M=OilA^@q
z)Rfe`<W&9e{G#mQ0w7mEIX@*;KPNvuv$!NPIld@0y(qP~I5R&lzMv>SD>b=9ztF%a
yJ|(dvF+MpbH8C$UFFignCr3X%J~J<~BtBlRpz;=nO>TZlX-=vg$l78cW&i->qcAxD

literal 0
HcmV?d00001

diff --git a/q02_data_cleaning_all/__pycache__/build.cpython-36.pyc b/q02_data_cleaning_all/__pycache__/build.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..ac010dcd50a5da4a7bfc65e118421a372a5d4d49
GIT binary patch
literal 1133
zcmaJ>OK%e~5Vkk_%95lhLeO%oxRgUu;EH&NB0xMYAb~89W!BSlgZEL}K?H41g(Lh7
zB>oL&Upe(JaDnkgpuz!L{_Gvk*fZaZ_u6oniLHk(Nl3_da;inRzJ<elMvz2OMc8_#
zi~^Yl^Bz#d@HEQcn65ir>qac@Y!jB^+*AEBV_CNssx#$)4Jd(lK9tchWg{5PH)Jf6
z;{ewK7<rjYW&Z<VV;K8yw4_hTZgzq`ktZ(Y{Gd?4`7U*1t3^?93&wKOsKTObqh455
zfaV&?`mIp!ccidls%la3MuRMJThtX_RJj6Cft=*#txF#H?FaiO6z|AA_}0e}<G6#v
zd_ic@m2EQrJ=0?%1KIl&po^i5d>G#KS-hg}F1K`_yxLg>Eop<bO*aQ<lWqU8JKB4W
zr_Fmv6!{n+ZhQ<5P=D1sB^`H?(1vZV4dp-%<>*uRGF(Ngc$KVDFKweXZj&}Wh>@zz
zCob~=-7f1zWuGA_Nl`UQ<iO&4P)X1%c`URpK(cHf8ZWA>VZ*1O4*1=&QS}nH#vLi^
zL-;e$B^%xG1X_N-Ft*l9?TON!p&1Tr-)^3G&UORVU+1KF4{o4ECF_zKD=eJ+#HP=9
zrO3sYNL2|JmY4~|T7nA;Rp+|r75Xmpsbgs_ELV7?qu}TZN0%;n!MoJE7#Em*mn?q^
zaDjraL~r34KiKrJE@6twIbe`EOrO{KjcM?6n&W*~7c)G$$gx~z8jSI^(dc>(xt$*E
zUFF^b-jzyG&A3o%`g&0)Icb(|_}AVhUT4M#B)CY&G^0E8JUvUd0(}u>yPK|Omc|9O
zabY9up^HW1iFb)CbR|k~af3VwRB5D(szr%61XaT>m^WCcx|(=;6gQyIHT1IM$lXS#
zm%`SiDqc_i6U*di?+X9tWlXwAH2wsLOEn0|{Zr_7FU5z3^y|9m-0#Rd1>2p?b#!T=
Y;D*nZIfpQ&muQN8KnFpFm`3T~7eRJM1ONa4

literal 0
HcmV?d00001

diff --git a/q02_data_cleaning_all/build.py b/q02_data_cleaning_all/build.py
index b56e2bc..56a42f1 100644
--- a/q02_data_cleaning_all/build.py
+++ b/q02_data_cleaning_all/build.py
@@ -1,3 +1,4 @@
+# %load q02_data_cleaning_all/build.py
 # Default Imports
 import sys, os
 sys.path.append(os.path.dirname(os.path.dirname(os.path.dirname('__file__'))))
@@ -12,3 +13,19 @@
 
 
 # Write your solution here :
+
+def data_cleaning(data):
+    
+    data['LoanAmount'].fillna(data['LoanAmount'].mean(),inplace = True)
+    cat_list = ['Gender','Married', 'Dependents', 'Self_Employed','Loan_Amount_Term','Credit_History']
+    for e in cat_list:
+        data[e].fillna(data[e].mode()[0],inplace = True)
+    X = data.iloc[:,:-1]
+    y = data.iloc[:,-1]
+    X_train, X_test, y_train, y_test = train_test_split(X,y, test_size = 0.25, random_state= 9)
+    
+    return X,y, X_train, X_test, y_train, y_test
+
+data_cleaning(loan_data)
+
+
diff --git a/q02_data_cleaning_all/tests/__pycache__/__init__.cpython-36.pyc b/q02_data_cleaning_all/tests/__pycache__/__init__.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..e116bcfd9f3c6e32eb5dc4ed4482fb0b1880648c
GIT binary patch
literal 182
zcmYL?F$w}P5Jj_KA;KQP1K2pCt;n`EHWoI*5GKPK&1Tk2EIg8j@FKR}!pdwRe)y07
zfAd#$J()hYM^p43@KaSnQ>>!`XZB`R?H(o#^B*6AlXMI@Sz^bL7!Hyy+LsC>jv+|v
zHG9IuZDLg~5&I-(rVQ7`5}nN!DX@(`wrE3u%$gMtee@Y|-rry2btmo?cr%=9vH1ce
Coid>S

literal 0
HcmV?d00001

diff --git a/q02_data_cleaning_all/tests/__pycache__/test_q02_data_cleaning.cpython-36.pyc b/q02_data_cleaning_all/tests/__pycache__/test_q02_data_cleaning.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..717ae059bf2855c5f050d3b8138dae6c130af0a6
GIT binary patch
literal 3404
zcmd6p&2HO95P-QPMN)tMi{r#i(y(bAKt(Dg?x8JE6e*H4Fc7#v)1tv50K;%s3RQ}f
zcb7&bq$lMm`UZW7zQ|sC@+<Vz8Pc|7+4WCyF$K;nm%Fn&A7?f5WwTkk_4n4xUuuT&
zk1=yO(0&NF{1XH-7)cCDK%0p<u&70_&r0k;g;t<#vr6I&T<RA6PT~!!v`P$4`ZeZG
z2(9yazrj3KotV(B@OshbqJHVnCS9P5bcrt06~5TNGBNtA{i_p$uJNUrr)y6QR%7*d
z2CH*xV$ti2JU3d6J=mYI-6GoinM?I3lpNZNS+0_pi$L%}wi_mUCJci9CJR*<L<tYm
zIPF2lLXWF%o+P2@$sv!%cW^)_$-*=k3eMt4#aSBUX)%Sfb|iPH`wU-iJ>DZ=G{WT%
zE?ngy-0~R+UI3#eBg}kfP>Wg21`Vi~1G>Uo=7DxtmDNDItj-#sJ=SCkpsQ?=ErG7F
zWpHyvFQa>pXSW~W`gq~89^8-MmYX1yp~!^%LEajDQ(12bBkRTlwgFq1dG4BP#-#X?
zuz<ZS(425VjBh{r9bhQVeBhD~c&eno!S4A_<6WLYt7Qv}4^aW3>!Flfs3$vlm}o2E
zY0K2M<VjcCNcAs6G-24kJDoS#fOnJ#BhGjl@y@TA_)QMsu@hyC1N2@jRUE-L>j^F;
zf**>k4;$(1+}{k2e7Yb^l8yo*vXH^C0qx;P-^2m)^9~$ZfK5pogkgDP+&)IPecI$B
z=+Q`l{$ps&qA$VYp*h5CZSPqyYW&S1VaB^*o+uxpH0RQ1nV)9LAA~A;<M#^QM@O`*
zDSJ@`TP;)6L5T(mL|qE>Pv1DfiG6eO3JfR#5CHJSxd0r@J@g-A21B9n--iZ-90Z>-
z`&QrXR~S(ahT=@1ncL$Zr_Mt>`)V}gem4`o;3^ksNy3k%|7sYfER;S*{SP7>@K^Wz
zQI<nc^8K-Ji2V)Oa`jRfgj;eel`4ej($0$jo!91&X*<dU*Oe~X(A7h8+Qy|@q~te;
zZsP=W<pu+s6_n;b0+%HMFvtQi$@r^tVLWc<vH=w<;F5O1NfE}WScNBX6~!8gYbdUR
z_#bw6&t`{a@x08!`}%$0p`pu3kG!}jH~%bdcjky~UYyt^Wi^_z(hlU4m`m+EFZ5}V
zb{4VG-47-<x}3z^(Xq368+;W$iVrxO`({6%<96}0xo_xl(sN*UkR7H$C(Z@$tfWv{
zZO^g0cv2|sY`5y#3xX7~RuJe~5DYSwC)jQT0hU|k8?lI=d>O?$iW?|yq4)&Frzkd1
zOfy{3!U{oP-H1<e34}+yisyRv!LNARTBlB#hznZU%hOmD@UZTr-$SuB3RN~p;@9m{
zYPj~!{RhEZS>Aq~#|aZy5$Q_VE=o!_PO%zSMWsVM!9x~c<y|cbb1Vd*j278Y+=Q6|
z(}=)CEU=^z_#Y`g13{fwrcm2!r>w2C6{aIv`>Cih0_dalY5PIp>q&^T_if4jJ1ll2
TCP2?@kn&dp?U7YT{tM(kzzm*1

literal 0
HcmV?d00001


From 4bd0136f29cd7d18fcc9602cc7462c6d17b1a6a8 Mon Sep 17 00:00:00 2001
From: tracedence <shoaibrashid887@gmail.com>
Date: Sun, 27 Jan 2019 04:36:25 +0000
Subject: [PATCH 3/4] Done

---
 .../__pycache__/__init__.cpython-36.pyc       | Bin 0 -> 178 bytes
 .../__pycache__/build.cpython-36.pyc          | Bin 0 -> 1862 bytes
 q02_data_cleaning_all_2/build.py              |  59 ++++++++++++++++++
 .../tests/__pycache__/__init__.cpython-36.pyc | Bin 0 -> 184 bytes
 .../q02_test_data_cleaning_2.cpython-36.pyc   | Bin 0 -> 4759 bytes
 5 files changed, 59 insertions(+)
 create mode 100644 q02_data_cleaning_all_2/__pycache__/__init__.cpython-36.pyc
 create mode 100644 q02_data_cleaning_all_2/__pycache__/build.cpython-36.pyc
 create mode 100644 q02_data_cleaning_all_2/tests/__pycache__/__init__.cpython-36.pyc
 create mode 100644 q02_data_cleaning_all_2/tests/__pycache__/q02_test_data_cleaning_2.cpython-36.pyc

diff --git a/q02_data_cleaning_all_2/__pycache__/__init__.cpython-36.pyc b/q02_data_cleaning_all_2/__pycache__/__init__.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..cf70d1328228e61590b6fa5b5a205dd48433b139
GIT binary patch
literal 178
zcmYL?F$zL45JdCALWF+=4`AcUw-vFqv9YibmXItlnxBo?Sa>84;YDn{g_Ze*I53Zy
zX=d5Blj(DNR8{W*KQ%RMs%2DR)n3h--9zPJ{^Nsi;+~*nZOBqI1P9I+NwtDuOAy$3
zDc)cvmr2BP#FY6=roeU4q18pB38bTs4s{6F0iut-AkNeMHD2{>Z-Lgsu~e%sM)5Gr

literal 0
HcmV?d00001

diff --git a/q02_data_cleaning_all_2/__pycache__/build.cpython-36.pyc b/q02_data_cleaning_all_2/__pycache__/build.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..ea9f5c9609a908a9f9a6e53e1671ba5049df582f
GIT binary patch
literal 1862
zcma)7+in{-5MAzFvSd558^>-Nq)A#pC}71l`W6HYk~An#8$q3-DPdqi?~LVDxtDNv
zlgLn>+*kiVkss(U@wHF+g+6t-zEm>opjvP^b2ywmb3|~ry4`mFyY0O{=Pm1RYwG8d
z|2>WJfDBl`q{Ue?+uWhqmTu~CkC`<``f0!eW+5Ae;2tyH!q%t_9{9&L`2n^X%ro<z
z&*ylDclkVD;ENDI_=m-B0Nb_tt&j9BYp2h27epnZSfa?1Y)HUDURN?f6)8;fH$ooW
zH5mh$i!3S>LK0U=o<()m>_J$M%Qrgs$t?f)^dqCPv1^LcFQ8$%dqSgpVO!S38c}~>
z5A9<%V(;CFHE~z1kzF|>>cvD4DF+<FJaA1{A8f<>#AXI>HMm)6&){<xcxMLhUf@&t
zl|Kro*8HkP9KvJky+^FfY}mCX9xT8jJvUBVgTUe{*{FTCt9P->Z?gT_?9m*|Ll2f{
z|MJiy{sDY;;u!9u4%~zlxCN^p975MlJ0kRR_yWFcSf66-l0s}oA&N*Lc8x;p8ig2K
ztq_Cj6=E|Aohrp9?2<x+GzzgZg&0gLWU{9fGW)Mnh+V0WK`$#r{ZH(PbKufBzP+QJ
zB&)P{K<BgK{&UO#l`^MLd1U%FcejN^?*D)(%`fSI2T4{)5hD+_g;EIuB{{!g$?ct-
zx1K>A3v<E=4WBCkwW#o@vy)eM&LiI4MR^cCOADEg5qS57$_rH0IC`v5XjkF^F-nNZ
zyS}5H!~Crd3>M{?4plNdG#+n1#lo=7s*-y@*Gfpu9`tSPig!t=LL#p_|D{A5d^*~k
ziMHr9Y(ulnBYW#xDiOrFtkdkY3SXI_7wux2)M{U3dTDmy7Xwn)t&0rqonl?xEm2}z
zMNo|kEY*@>=ZT__V{K=JcFWhQ(t*N8SZx=e=Y~k>b($tv>N$lag^eWIOHmC`wM`8d
zNILB+%tVU)KnJg)N{J-XJ~>!Y&G9)v4goyK>NKMM$icUCA?3Vl95@&AFH7C0%Kz*Q
z4)YWTR4qmlImW@Sx%#~<C>+E&AYEv~q^y#d?zth+E6t@w<wn%z;Pr#eXy#swgp|?d
zV82cztQTXwFuS_BK(8zxk+JMWw#;swJv|n(9&6L+u{HZ9TXsIZs=lEL{hsy<k%1`H
z8im@Uj*GEcG5!;xIu=!)%4C226CJEy*};0lU(rQu>Bs-S(nfUYlW<nYF?B*Z@(`JC
zK&$yg7-7_Hv%+3!Hdc4YRbLr%*O+_8m@b=tY07*}m<}JNIn)v#nbPGVnUEChu#Vj(
I8?t5hU;mH(0{{R3

literal 0
HcmV?d00001

diff --git a/q02_data_cleaning_all_2/build.py b/q02_data_cleaning_all_2/build.py
index e20ff7b..23a901d 100644
--- a/q02_data_cleaning_all_2/build.py
+++ b/q02_data_cleaning_all_2/build.py
@@ -1,3 +1,4 @@
+# %load q02_data_cleaning_all_2/build.py
 # Default Imports
 import pandas as pd
 import numpy as np
@@ -11,3 +12,61 @@
 
 
 # Write your solution here :
+def data_cleaning_2(X_train, X_test, y_train, y_test):
+    
+    num_col = X_train.select_dtypes(['int','float']).columns
+    X_train[num_col] = X_train[num_col].apply(np.sqrt)
+    X_train['Gender'].replace({'Male':0, 'Female':1},inplace = True)
+    X_train['Married'].replace({'Yes':1, 'No': 0}, inplace= True)
+    X_train['Education'].replace({'Graduate':1, 'Not Graduate':0}, inplace = True)
+    X_train['Self_Employed'].replace({'Yes':1, 'No':0},inplace = True)
+    new_pro = pd.get_dummies(X_train['Property_Area']).reset_index()
+    X_train  = X_train.reset_index()
+    X_train = X_train.merge(new_pro, how='left', left_on = 'index', right_on = 'index')
+    X_train.index = X_train['index']
+    new_dependent = pd.get_dummies(X_train['Dependents']).reset_index()
+    X_train = X_train.merge(new_dependent, how = 'left', left_on = 'index' ,right_on = 'index')
+    X_train.drop(['Property_Area', 'Dependents', 'Rural','index', '0'], axis = 1, inplace = True)
+    
+    X_train.rename(columns={'1':'Dependents_1', '2':'Dependents_2','3+':'Dependents_3'}, inplace = True)
+    X_train.rename(columns={'Semiurban':'Property_Area_Semiurban','Urban':'Property_Area_Urban'}, inplace = True)
+    
+    
+    num_col = X_test.select_dtypes(['int','float']).columns
+    X_test[num_col] = X_test[num_col].apply(np.sqrt)
+    X_test['Gender'].replace({'Male':0, 'Female':1},inplace = True)
+    X_test['Married'].replace({'Yes':1, 'No': 0}, inplace= True)
+    X_test['Education'].replace({'Graduate':1, 'Not Graduate':0}, inplace = True)
+    X_test['Self_Employed'].replace({'Yes':1, 'No':0},inplace = True)
+    new_pro = pd.get_dummies(X_test['Property_Area']).reset_index()
+    X_test  = X_test.reset_index()
+    X_test = X_test.merge(new_pro, how='left', left_on = 'index', right_on = 'index')
+    X_test.index = X_test['index']
+    new_dependent = pd.get_dummies(X_test['Dependents']).reset_index()
+    X_test = X_test.merge(new_dependent, how = 'left', left_on = 'index' ,right_on = 'index')
+    X_test.drop(['Property_Area', 'Dependents', 'Rural','index', '0'], axis = 1, inplace = True)
+    
+    X_test.rename(columns={'1':'Dependents_1', '2':'Dependents_2','3+':'Dependents_3'}, inplace = True)
+    X_test.rename(columns={'Semiurban':'Property_Area_Semiurban','Urban':'Property_Area_Urban'}, inplace = True)
+
+
+    return X_train, X_test, y_train, y_test
+
+# num_col = X_test.select_dtypes(['int','float']).columns
+# X_test[num_col] = X_test[num_col].apply(np.sqrt)
+# X_test['Gender'].replace({'Male':0, 'Female':1},inplace = True)
+# X_test['Married'].replace({'Yes':1, 'No': 0}, inplace= True)
+# X_test['Education'].replace({'Graduate':1, 'Not Graduate':0}, inplace = True)
+# X_test['Self_Employed'].replace({'Yes':1, 'No':0},inplace = True)
+# new_pro = pd.get_dummies(X_test['Property_Area']).reset_index()
+# X_test  = X_test.reset_index()
+# X_test = X_test.merge(new_pro, how='left', left_on = 'index', right_on = 'index')
+# X_test.index = X_test['index']
+# new_dependent = pd.get_dummies(X_test['Dependents']).reset_index()
+# X_test = X_test.merge(new_dependent, how = 'left', left_on = 'index' ,right_on = 'index')
+# X_test.drop(['Property_Area', 'Dependents', 'Rural','index', '0'], axis = 1, inplace = True)
+
+
+
+
+
diff --git a/q02_data_cleaning_all_2/tests/__pycache__/__init__.cpython-36.pyc b/q02_data_cleaning_all_2/tests/__pycache__/__init__.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..a84e1c4b923bcdf344f6243e2180ade2142c5f78
GIT binary patch
literal 184
zcmYL?F$w}P5Jj_KA;KQP1K2pSwj#DRHWoI*5RzeyW;1JMEIg8j@FKR}!pdwRe)y07
zfAg1hJ()hYM^p43@KaU7rdUP=-t5h++C5Aj=083JC+-L|TSuOfBRFurNM9;A_5_jJ
zAQ~KIYBPxhPS|HYlT)}Z8uV5zx=1#K)S`_M8&DDzKny7;;=I4V#;cC~Ez)|pmtyk;
Di&`@@

literal 0
HcmV?d00001

diff --git a/q02_data_cleaning_all_2/tests/__pycache__/q02_test_data_cleaning_2.cpython-36.pyc b/q02_data_cleaning_all_2/tests/__pycache__/q02_test_data_cleaning_2.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..66819a11eab2e76926d2d76ec01baf7f10a3a50b
GIT binary patch
literal 4759
zcmd6rOOw+^5P<cvEI;?Xun<5Y0pb*~@sbB2kb>e>Ta|KBAt@*oRYG`XjT~7rGZGcM
z$jMyt4|2*qKO+Aj=eg$OzmQY9CEF`+yw2Kiz%KQ)$D`KFH$B}mx>_pb&-}Uf<oA*!
z{Vk<FI_S6H;(vi45|KU0lA%;QrK4J^ED@Dxp4Q1&8PFw?@$^pC%Er3x8J(P!i}kFR
z?-Z<pEK#{#B*s9tO0?82lN`wp6wq_D6xU_3e4<+`)~dB;ty>$`Ce6?_x=uIf=0JO*
zO={OjVMl6j4J2!uu94zMKS9?g`bp5oZFU|=q(sWEB~qs9K($U0`Jq%<8N#TfMnx9J
z50r;@98QM{)Ih}waYDy#dere<zh&=2RcWdU+XPYQxs=(Ac7m6V2UX=34fi9@b6AUa
zX|sPG@2Po#<J(<EiQ5d_z_%kmHcUyi$zNL8Z}9KEyF(csZ58j*F7*lZLvGitjqg~{
zr7Y~(w;6TpAK5d<Z(^s%AMEZexVQ&`#*<?yL?+5>$x?B0pwvi)=ul=zmKad#h#kr-
zDUc$RMuNu5krlEEzw=~`tV3BK8)Os8BH1F_P?pFEauUii*#UT`#0EmUJI$$ygTN1V
z%z*0-TznTqD24Js{zZNzwUtnPB@=l^8bC*(ld0eP%C<B>zmnh)T7#jvp$4LV?%vNp
zG^E4?xKRfInHR`q^MU)4`k+-bmW6f<8;gR&Ibgf@B62*TdepBdLgUoi7aC5@9}-)E
zX8%OB=Rt?oLgq9nAaByzqk#RyyYREt3<!mpYq>mhn=r*KMmfjX>$0E?9jU#zvTIMd
zgzb2qy;}=`3XlIq?NF25YPToO&icci3~K|r#&H;uYRLWF8I-H@4NOe}8j2_3EGX$o
zNTvW6^8~os7^={!|K)@=q5Dqcg(kRnM7c=<(+@(k<AlxUW-BH?-J+r%%mOQbt%|}R
zs7S1g0-+~j6*O^X<}gME`3^KmFvx(Rerycm`zI98hcJ-E6iiUS1;H6ZrLDHLc818I
zj;@(<Jf+e9_5cBRxEc1k)Z7o4$!HicKjC4z+<e+~eBy8u*TH?}bm-H|W-o}qN$IF9
z#1HcVuVlq);*@)Q&*vemV=DB=HX;y8mk6yHFe)<pXhY;C=7ffiuE+_+LjW($1ekQu
z#{2B#0q0=Q91%!zSytr!SI36=Ut8}QR<;Sl5?L51b6lTo!!LFM#Yq(CMC=rZBM5$U
z9D->0L$HxXcw<?FXzU$j*gMQn=nzF+%7y+gF5isg$b99kkQ^C%H%aQyJ6P8#XT%lA
zOcZmR@&A`|Vz+vT;Bw13u{8y7k}I&2<0uJ98q#zN67?icXbtxfMAhMKXx6{>eL5sI
zg9tM-2_Rb>qzXFD-Ekf<lSYs$FhdB<gB&Yz@%pvnebnN?F1DMv@25xyw;$3Hfi7id
zSu$HlW&n?WV>0XlSP}UNnEEn7EH$F~7hrFUX5j#qHnb$0;Unn45O;%biH#A5Z|K<(
zWG>DN@1p_1LcN~ia(&Ujeqd%<Dh#P$&U#B!a$_n-In3NZb1<Iz%@mh~gG|cKY%rL(
zlOWSph?WX3na1!pTtv8M>|dD*FZ{U!8?m|31$LZQ;YNmWU=(85d!UO<eDe{WQkmx|
z@r}5b!sLAmY>gMu%|)GM%F^OkeuZbp16v9UF8?VkTUhZKh$EoP-5pJa0t+bKT90Dx
z!ht1is8LxFhHd+>3ASyKx9v_qA`kVVZR2S!`G>7x&+(wdPNO)3;(Zh!ptywMV-%Pw
z83r8oIf`p24mQ!ThDQ;W1F;Sl$6cc#8#;V5@mDw0@i%T&Ic@35o<C0S!ian~MCV~R
zevx?=xgLp6q1GKZGxnS?=y>k4>YSrwb>>mBnpnz!RqOWgE!FKquO&uyaQiQ^u8;2}
z!uXJ88SqYl*!T`27r&yw(=5C#fL9Jeek|l(ylhzPBzSma=dmlrWL+mIoL$6v+|jX5
kP}ET%%UO_xaU+?{uVHaZ98DT>ar{;Cpc(R}ygjM;7e{X+8UO$Q

literal 0
HcmV?d00001


From a22759bf4d2b74caa0a61e52f52e3f5a37533320 Mon Sep 17 00:00:00 2001
From: tracedence <shoaibrashid887@gmail.com>
Date: Sun, 27 Jan 2019 05:22:03 +0000
Subject: [PATCH 4/4] Done

---
 .../__pycache__/__init__.cpython-36.pyc        | Bin 0 -> 178 bytes
 .../__pycache__/build.cpython-36.pyc           | Bin 0 -> 1264 bytes
 q03_logistic_regression/build.py               |  17 +++++++++++++++++
 .../tests/__pycache__/__init__.cpython-36.pyc  | Bin 0 -> 184 bytes
 ...test_q03_logistic_regression.cpython-36.pyc | Bin 0 -> 2298 bytes
 5 files changed, 17 insertions(+)
 create mode 100644 q03_logistic_regression/__pycache__/__init__.cpython-36.pyc
 create mode 100644 q03_logistic_regression/__pycache__/build.cpython-36.pyc
 create mode 100644 q03_logistic_regression/tests/__pycache__/__init__.cpython-36.pyc
 create mode 100644 q03_logistic_regression/tests/__pycache__/test_q03_logistic_regression.cpython-36.pyc

diff --git a/q03_logistic_regression/__pycache__/__init__.cpython-36.pyc b/q03_logistic_regression/__pycache__/__init__.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..1871534aa0713d3f8dd587acca40c00950e88fdf
GIT binary patch
literal 178
zcmXr!<>ivy>lwoU1dl-k3@`#24nSPY0whuxf*CX!{Z=v*frJsnuXz28{M=OilA^@q
z)Rfe`<W&9e{G#mQ0w7mEIX@*;KPNvuv$!NPIld@0y(qP~I5R&lzMv>SD>b=9ztF%q
j9*ewwe0*kJW=VX!UP0w84x8Nkl+v73JCMc2K+FIDUF<Pv

literal 0
HcmV?d00001

diff --git a/q03_logistic_regression/__pycache__/build.cpython-36.pyc b/q03_logistic_regression/__pycache__/build.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..7f3d2cb2e3aa58a43914d7bbc13b187d332081b5
GIT binary patch
literal 1264
zcmbtUJ8#=C5GG~4Y)5utJ5IWG^J2Hj8uWp*LxD6yGc+I!2#QF%a;Qg0Ij9SHsrys9
z^e=HOI^-{O>XEUFq|H#E5_mj)clvZM>M#fzKc2rGe|8Z1j_zD8=r3T&uOKL*m?MH=
zx44zoNDXVO#_iN0j?r!IrXKN(?r=Y?le*De-bkCIX>^Zor2z>rVtCe~{tA;eYtOc6
zoi<h$=yldMSX1v~m+X)Z>5?AVC4CmKF6*&f)?e9UdxO#@-TH#a9>r&95Zpkd=yZTp
z>#U428VPz9N1O>jIw$!wk!2FUV^hJTO!5rS?KsaSbNvveQ7Mv-fVJ~^$rC0*!P5LX
z;s6CSDx)ywEXtB>s(tPO;lZcJ8W{0B%ECf0n#85%oM+}8Mu)MyR^Ev=zj<|oA<Ed+
zvBFh{RfBa5OBNtXv_ea}LNiS96tD2iS~`_It5v9S9-$T1`w^<}Uv|8<!+wNH8(7@R
zojKrdtt^Z{cPrpqdZky@DsN>i{mO^3S*OaGK<MSb7ZwDd0?9aw%aE3ff=T7YIiIJQ
zRKX-E!%{?<oa7=EI%-t|OnK{^g-&WvSKhgHOETqx!X!8>HUKC95LcSU5HFjjA@GwH
zU-u_WbIQ1K7p9QPiqmgM^nv&1c=RDp*$9|pMp+iK(feFn$^sroF}wujzilyeZqs8R
z$!Czu=<3N~_<((MF;6%h77NvVz}kh+${q-0xmF8z%yRJ-Zon<}0Uzuux6q>|l~qu(
zD=%S4L^gziz?Wk^XxY?MmGm|?o@C&2o2MxvX>Uv^gCUHisiS%cNi3o=PkC}Nd?30_
zZo{i5Ps2Zkc4+bvaHf$u{&${(@cv*&oUfVm-$1?oFQ)Kd&E;t>U$BUV`drk_vC$t^
zHWhiH@VUYZu?=2gM~jXYx)-M8W*F{>JuoQ$2*zg4*)u(qQg_COHvPV>2IxNS;lBL~
D=X-Y&

literal 0
HcmV?d00001

diff --git a/q03_logistic_regression/build.py b/q03_logistic_regression/build.py
index cdbd506..b93fc1d 100644
--- a/q03_logistic_regression/build.py
+++ b/q03_logistic_regression/build.py
@@ -1,3 +1,4 @@
+# %load q03_logistic_regression/build.py
 # Default Imports
 import pandas as pd
 from sklearn.preprocessing import StandardScaler
@@ -15,4 +16,20 @@
 
 
 # Write your solution code here:
+def logistic_regression(X_train, X_test, y_train, y_test):
+    
+    scaler = StandardScaler()
+    #feature scaling on train
+    num_col = X_train.select_dtypes(['float']).columns
+    X_train[num_col] = scaler.fit_transform(X_train[num_col])
+    #feature scaling on test
+    num_col = X_test.select_dtypes(['float']).columns
+    X_test[num_col] = scaler.fit_transform(X_test[num_col])
+
+    logistic_model = LogisticRegression()
+    logistic_model.fit(X_train, y_train)
+    y_pred = logistic_model.predict(X_test)
+    cm = confusion_matrix(y_test, y_pred)
+    return cm
+
 
diff --git a/q03_logistic_regression/tests/__pycache__/__init__.cpython-36.pyc b/q03_logistic_regression/tests/__pycache__/__init__.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..2b945b421243c3f04f5e87f7f20d1014f9cac9df
GIT binary patch
literal 184
zcmXr!<>ivy>lwoU1dl-k3@`#24nSPY0whuxf*CX!{Z=v*frJsnuT=ev{M=OilA^@q
z)Rfe`<W&9e{G#mQ0w7mEIX@*;KPNvuv$!NPIld@0y(qP~I5R&lzMv>SD>b=9ztF%q
p9*ewwNosLPv3`7fW?p7Ve7s&k<t+}I-29Z%oK!oI^~FHU005=^GtdA4

literal 0
HcmV?d00001

diff --git a/q03_logistic_regression/tests/__pycache__/test_q03_logistic_regression.cpython-36.pyc b/q03_logistic_regression/tests/__pycache__/test_q03_logistic_regression.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..6dd09a48f61e60c16c69b37c6d042e9f95cad98b
GIT binary patch
literal 2298
zcmbtVPjA~c6c;5~Rvg<&+ca5*{m}yCARw^Y^wgmnT4X^HAjo9{RuI8hRb*Uwl;n_9
zP#f*Z{WALyyX?AepzBWg3OnsRDRJyHLDx+JAD<qnN50?t@$pfo)B5uFgX2Fv%lgZj
zd*Iu83Qhk3f?J$Oi;}9hsSUj?oz$f+!7)cRQjdC-?#gEBQ@_$1aw%=m7U&*t%H_08
z+m+pyowQ54#1dq<!k4Cmu8P&+9p2*0QycV_SgmXx_QxK*OYhM&dY`V-4dIJ5abK*9
zjj4O=&aCY#EB)-b#oN5|-r^nMP93_*$&XgpJ%>1~qmY#TPePl$NDI(AJTeiBrHHa5
z8-QVD!N7LWlov)OLNO)M{4A2^HaG(NJ(&*@ZIT!U21;n1<QbT|17S`IDI+z|BN0!w
z&|_cbQN~6}@FX@cj}=+9sNlU=pV8(a4j=q*PQZ1H8@+sRt*6lR4hT`D0N^Lwes58S
zJKTld<qhsZ-{4K|L+|k=2)I>lV4#N|ZZgJI@ImK3v^{8g7lg44nUZ(ptu?fb^OkV3
zX-#1<SZMD1#9p_i4rUUTagP+bJ9j}$zTSTWAsfL1h`vZgW^}O4cY^21naDs3U4`qZ
zCWvx5(po69e_BMcbfm~ayL7dXC#8$n{bs3EnD+O$|2j`a->4`STx7B6|D3B|^$0%u
zan1!m`#~B2>3j$q>7PD$#BLbx8$d=^lHD?;HyW2)IPIqY*(|)futklrTaHgA-`;|F
z@8gaZl7bGaq<jU+d{XL^;PIO!z&Sd1VAkaE40|q4qC%PgxLODu@I1(J6Qq%eUk8IK
z6N@uK=PE0z4ZcEKb)i!$C@=}NxC2`G&8?waIMo|4t%igEB#%ECl1o1`M37%tjw(dH
z00rm?L^bx%;l#K|**QYmN0Y-VDB$b`W;_zXNv;AVOrf$Gmmtx>%dAL8<6g!irK0i6
zonV|7z+7=L5h_Bkt;0sSS~K-PA7t7@K*7?E)3Q<hO`8|LmoBaok{Y^esD-zpR_}>b
zWd|K#d>W&*NQ>BH^6ZoGyY6uT#~HKMXE+Y+DU|m`M*RY#lZP`H7l7kAkO~=9b7_7I
zn1uXM=)82&=*=x~Y<~=nin?3iMPrS!$mF~K3Pa`TC~THKV_B36#>y6BY0e9YdYdt<
zKlO;h)2!A|Y@)zi%<;d<DINmV0)dLgYUUHa0k2nC!_TQVihvTjv`1XwSt$J?OAJP=
zFc`WAm&G_Tc`B1vy&H<tW%25rKG>~!ad|(nNXp);B9UC-k*E9r%M;tZ@P(Y9&;Gry
zhwR#oc;Tmk!8YpsRTaLOWOxOds=QENiHI}2?v|?C5GfCLm&!*ac~O!vZN6Yo<&%th
tpa5L7IleT2g7_h?DwuD|O6fnT)A~KatTB0z8q%J7TcG=7jjWTd{SP6Omi7Pu

literal 0
HcmV?d00001