import pandas as pd import matplotlib.pyplot as plt import seaborn as sns data = pd.read_csv(‘telecom churn.csv') data 7038 7039 7040 7041 7042 Customer ID 002- ORFBO 003. MIKNFE 004- TLL 0011- IGKFF 0013, EXCHZ 9987. wuivo 9992 2RAMN 9992- JOEL 9993- “HIEB 9995 HOTOH Gender Age Married Female Male Male Male Female Female Male Male Male Male 7043 rows x 38 columns 37 46 50 78 20 40 22 21 36 Yes No No Yes Yes No No Yes — EEE data.head(10) Number of Dependents o o ° City Frazier Park Glendale Costa Mesa Martinez Camarille la Mesa Riverbank. Elk Solana Beach Sierra City Zip Code 93225 91206 92627 94553 93010 9194" 95367 95432 92075 96125 Latitude 34,827662 34,162515 33645672 38.014457 34,227846 32759327 37.734971 39,108252 33.0181 39,600598 Longitude -118999073 -118 203869 “117922613 -122.115432 -119.079903 -116997260 -120.95427" -123,64512" -117.263628 -120,636358Customer ID 002- ORFBO 0003- MIKNFE o004- TLHL oo11- IGKFF 0013- EXCHZ 0013 MHZWE 0013- SMEOE 0014. aMAQL 0015- yoco 0016. quis Gender Age Married Female Male Male Male Female Female Female Male Female Female 10 rows x 38 columns 37 46 50 Yes Ne Ne Yes Yes Ne Yes Yes Ne Yes Number of Dependents 0 0 0 0 —EEEEEEEE data.tail(10) City Frazier Park Glendale Costa Mesa Martinez Camarillo Midpines Lompoc Napa Simi Valley Sheridan Zip code 93225 91206 92627 94552 93010 95345 93437 94558 93062 95681 Latitude 34,827662 34,162515 33645672 38.014457 34,227846 37.581496 34757477 38.489789 34,296813 38.984756 Longitude 118999073 -118.203869 “117922613 122.115432 -119,079903 -119.972762 120550507 122270110 -118,685703 -121.345074 Num7033 7034 7035 7036 7037 7038 7039 7040 7041 7042 Customer ID 9975- SKRNR 9978- HYCIN 9979 RGMZT 9985- wwix 9986- BONCE 9987. wuTyD 9992- 2RAMN 9992- JOEL 9993 “HIEB 9995 HOTOH Gender Male Male Female Female Female Female Male Male Male Male 10 rows x 38 columns ‘Age Married 24 No mR Yes 20 No 53 No 36 No 20 No 40 Yes. 22 No 2 Yes 36 Yes —EEEEEEE Find the shape of dataset fn data.shape (7043, 38) Number of city Dependents © Sierraville 1 Bakersfield Angeles © due © Fallbrook la Mesa © Riverbank o elk ome Sierra City 1 print("Nunber of Rows" data.shape[@}) print(*Nunber of Columns" data. shape[1]) Number of Rows 7042 Number of Columns 38 fn data. info() Zip code 96126 93301 0022 93628 92028 91941 98367 95432 92075 96125 Latitude 39559708 35,383937 34,023810 36807595 33,362575 32,759327 37,734971 39,108252 3.001813 39,600599 Longitude 120345639 -119.020428 118156582 -118.901544 117299644 -116.997260 12098427" -123,64512" 117.263628 -120.636358 RangeIndex: 7843 entries, @ to 7042 Data columns (total 38 columns): # Column’ Non-Null Count type @ Customer 10 7043 non-null object 1 Gender 7043 non-null object 2 Age 7043 non-null int64 3 Married 7043 non-null object 4 Number of Dependents 7e43 non-null int64 5 City 7043 non-null object 6 Zip Code 7043 non-null —int6a 7 Latitude 7@43 non-null floate4 8 Longitude 7@43 non-null floated 9 Number of Referrals 7043 non-null inte4 10 Tenure in Months 7043 non-null int6a 11 offer 7843 non-null object 12 Phone Service 7843 non-null object 13 Avg Monthly Long Distance Charges 6361 non-null floated 14 Multiple Lines 6361 non-null object 45. Internet Service 7043 non-null object, 16 Internet Type 5517 non-null object 17 Avg Monthly GB Download 5517 non-null floate4 48 Online Security 5517 non-null object 19 Online Backup 5517 non-null object 20 Device Protection Plan 5517 non-null object 21 Premium Tech Support 5517 non-null object 22. Streaming TV 5517 non-null object 23. Streaming Movies 5517 non-null object 24 Streaming Music 5517 non-null object 25. Unlimited Data 5517 non-null object 26 Contract 7843 non-null object 27. Paperless Billing 7843 non-null object 28 Payment Method 7043 non-null object 29. Monthly Charge 7043 non-null float64 30 Total Charges 7843 non-null floate4 31 Total Refunds 7043 non-null floate4 32. Total Extra Data Charges 7043 non-null inte 33 Total Long Distance Charges 7043 non-null float6a 34 Total Revenue 7@43 non-null floate4 35. Customer Status 7043 non-null object 36 Churn Category 1869 non-null object 37. Churn Reason 1869 non-null object dtypes: Floate4(9), intea(6), object(23) memory usage: 2.0+ ME data.colunns Index(['Customer 10", ‘Gender’, ‘Age’, ‘Married’, ‘Number of Dependents’, ‘City’, ‘Zip Code’, ‘Latitude’, ‘Longitude’, ‘Number of Referrals’, “Tenure in Months’, ‘Offer’, "Phone Service’, ‘avg Monthly Long Distance Charges’, ‘Multiple Lines’, "Internet Service’, ‘Internet Type’, ‘Avg Monthly GB Download’, ‘Online Security’, ‘Online Backup’, ‘Device Protection Plan’, ‘Premium Tech Support’, ‘Streaming TV’, ‘Streaming Movies’, "Streaming Music’, ‘Unlimited Data’, ‘Contract’, ‘Paperless Billing", "Payment Method", ‘Monthly Charge’, ‘Total Charges’, ‘Total Refunds", ‘Total Extra Data Charges’, ‘Total Long Distance Charges", Total Revenue’, ‘Customer Status’, ‘Churn Category’, ‘Churn Reason" ], dtype=' object} Descriptive Statisticsprint (data.describe()) count. mean std min 25% 50% 75% max count mean std min 25% 50% 75% max count. mean std min 25% 50% 75% max count mean std min 25% 50% 75% max count. mean std min 25% 50% 75% max Age Number of Dependents Total Refunds Zip Code 7e43.000000 468692 93486.070567 1856. 767505 980000 ©90001.000000 020008 92101.000000 020088 93518,000000 980000 © 95329,000000 000000 96150.000000 7043 .000 32.386 24.542 1.000 9.000 29.000 55.000 72.000 962182 -902614 000000 000000 200000 000000 Total Revenue 220000 379056 204542 360000 610080 642000 145000 7043000000 [email protected] 46.509726 @ 16.750352 962802 19.000008 e 32.000000 e 46.000000 6 60.000000 e. 80.000000 9 Longitude Number of Referrals 7043.00000 7043.000000 -119.756684 1.951867 2.154425 3.001199 -124.301372 0.200000 -121.788098 0.000000 -119.595293 2.000000 -117.969795 3.000000 -114.192901 11.080000 ‘Avg Monthly Long Distance Charges 6361000000 25.420517 14200374 1.010008 13.050000 25.690008 37680000 49.990000 Nonthly Charge Total Charges 7043.000000 © 7043.000008 ©7843. 000000 63.596131 2280..381264 31.204743 — 2266.226462 -10.000008 18.800000 30.400000 —_400.150000 [email protected] —1394.556000 89.750008 3786.600000 118.750000 8684.800000 © 49.790800 Total Long Distance Charges 7e43.000000 7043. 749,099262 3034. 846.660055 2865. @.00000@ = 21. 70.545008 605. 401.440000 2108. 1191.160000 4801. 3564.720000 11979. data-isnull()-sum() #check null values 340000 Latitude \ 7043.000000 36.197455 2.468929 32.555828 3.990646 36.205465 38.161321 41.962127 Tenure in Months \ 220 767 261, 1200 200 1200 1200 1200 5517 000000 ‘Avg Monthly G8 Download \ 26.189958 19. 586585 2.000000 13.000000 21.000000 30.000000 85.000000 Total Extra Data Charges [email protected] 6.860713 25.104978 0.200000 2.200000 2.200000 2.200000 150.000000 \Customer ID Gender Age Narried Number of Dependents city Zip Code Latitude Longitude Number of Referrals Tenure in Months Offer Phone Service Avg Monthly Long Distance Charges Multiple Lines Internet Service Internet Type Avg Monthly GB Download Online Security Online Backup Device Protection Plan Premium Tech Support Streaming TV Streaming Movies Streaming Music Unlimited pata Contract Paperless Billing Payment Method Nonthly Charge Total Charges Total Refunds Total Extra Data Charges Total Long Distance Charges Total Revenue Customer Status Churn Category Churn Reason dtype: intes sns.heatmap(data.isnul1()) wheatmap visual for null values data[ ‘Churn Category’ ].isnul1( punne 7038 7039 7040 7041 7042 Name -10 -08 06 04 oz hoo Zip Code Longitude “Enure in Months ine Security Multiple Lines Device Protect ‘Streaming TV Payment Method “otal Charges “Btal Extra Data Charges “etal Revenue Churn Category Number of Dependents True True True False False False True False True True True Churn Category, Length: 7843, dtype: bool ‘#Renove missing values data. dropna(ho "any", inplact True) #Renove duplicate dup=data.duplicated() .any() data. describe (includ all’)count unique top freq, ‘mean std min 25% 50% 75% max Customer Ip Sender 1586 1586 1586 2 9004- , Tieu Female 1 803 NaN NaN NeN NaN NeN NaN NeN NaN NeN NaN NaN NaN NaN NaN 11 rows x 38 columns Age 1586,000000 NaN Nat Naty 50165826 17.684679 119.000000 35,000000 50,000000 65.750000 80,000000 —EEEEEEEEEE data[ ‘Customer ID'].unique() array(['0004-TLHLJ', "@011-IGKFF', *013-EXCHZ", ... Married 1586 2 Ne 1007 NaN NaN. NaN’ NaN. NaN. NaN. NaN Number of Dependents 1586,000000 NaN NaN Naty 0.092686 0.455859 0.000000 0.000000 0.000000 0.000000 5.000000 “9985-MAVIX', '9992-RRAMN'], dtype=object) data[ ‘Customer Status’ ].value_counts() Churned 1586 Name: Customer Status, dtype: int64 City 1586 598, San Diego 161 NaN NaN NaN NaN NaN NaN NaN Zip Code 1586,000000 Nan’ NaN NaN 93433361917 1830004432 90001.000000 92117,000000 93278.500000 95309.000000 96150.000000 » '9965-YOKZB" Latitude 11586.000000 NaN. Nan. NaN’ 36,009995 2.509295 32555828 33874789 35,363810 38,066743 41962127 data[['Tenure in Months", ‘Monthly Charge’, ‘Total Revenue’]].describe().T count mean std Tenure in Monthe 1586C 18902900 19:759294 Monthly Charge 15660 81.108670 19814937 Total Revenue 1586.0 2221295648 2543.063161 plt. Figure(Figsize=(6, 6)) data[‘Tenure in Months'].hist() min 25% toc 3.00 9.00 71.00 50% 75% 11.000 31,0000 83.925 95,3500 4692 26028 1142.35 3414.0875 max 72.00 11835 11195.44In [22]: out[22] In [23] out[23]: plt. figure(figsize=(6, 6)) data[ ‘Monthly Charge’ ].hist() data[ ‘Total Revenue’ ]-hist() 8 8 $ 8 $ 8 8 8 2000 © 4000» 6000-8000 10000) UNIVARIATE ANALYSIS Here, in this analysis will observe the influence of various features on the effect of churning In [24] out [24]: In [58]: data[ ‘Customer Status’ ].describe() fdistribution of customer status count. 1586 unique 1 top Churnec freq 1586 Name: Customer Status, dtype: object data[‘Churn Category" ]-hist() #histogram chart for Churn Category 828888 8 0 Competitor Dissatisfaction ther ‘Attitude Price Customer Churn on basis of gender plt. figure(figsize=(7,7)) sns.countplot(x="Custoner Status’ ,hue='Gender’ ,data=data) pit. show()800 700 600 200 100 ‘Chumed ‘Customer Status from the above distribution, it can be seen that their is a slight difference between male and female customers for churning In [68]: plt. figure (figsize=(10,7)) sns.countplot (x=' Customer Status’ ,hue='Churn Category’ ,data=data) ut[ee): 10000 ‘5000 6000 “btal Revenue 4000 2000 Monthly Charge # Sample data for Online Backup and Online Payment categories = ['Online Backup’, ‘Online Payment] values = [7, 30] # You can adjust these values based on your data plt.pie(values, labels=categories, autopct='%1.1f%%', startangle=98, colors=['skyb! plt.axis( equal’) plt.title(‘Usage Distribution: Online Backup vs. Online Payment" ) plt.show() Usage Distribution: Online Backup vs. Online Payment Online Backup corr_natrix = data.corr() ‘# Plotting a heatmap of the correlation matrix pit. figure(figsize=(12, 1€)) sns.heatmap(corr_matrix, annot=True, cmap="coolwarm', fmt=".2f', linewidths=0.5) plt.title(‘Correlation matrix’) pit. show()