mini project1 team5

14
Inspire…Educate…Transform. The best place for students to learn Applied Engineering http://www.insofe.edu.in Team 5 Mini Project - 1 Loyalty Program Satyasheel Radhika Rani K Krishna Chaitanya Saikumar Allaka Prashanth Soma November 25, 2013

Upload: saikumar-allaka

Post on 16-May-2015

255 views

Category:

Business


1 download

DESCRIPTION

We shall identify most likely loyal customer for the gaming company based on its customers details and their previous game preferences.

TRANSCRIPT

Page 1: Mini project1 team5

Inspire…Educate…Transform.

The best place for students to learn Applied Engineering http://www.insofe.edu.in

Team 5

Mini Project - 1

Loyalty Program

SatyasheelRadhika RaniK Krishna Chaitanya

Saikumar AllakaPrashanth Soma

November 25, 2013

Page 2: Mini project1 team5

The best place for students to learn Applied Engineering 2 http://www.insofe.edu.in

IntroductionWe shall identify most likely loyal customer for the gaming company based on its customers details and their previous game preferences.

We assume that customers demographic and game-play preferences would help us define loyal customers.

We will use the customers demographic and their past game preferences to discover relationship between the known characteristics of a customer to their reaction to the loyalty program.

Page 3: Mini project1 team5

The best place for students to learn Applied Engineering 3 http://www.insofe.edu.in

Business use

Gaming company can maximize their games sales and retain customers without compromising on revenues

Provides customer’s gaming pattern insights and helps gaming company in building and promoting right games to the right customers.

Page 4: Mini project1 team5

The best place for students to learn Applied Engineering 4 http://www.insofe.edu.in

Process• Derive new attributes from the existing data to get a

bigger picture and to learn their characteristics.

• Bin them according to the  Recency, Frequency, Monetary Value (RFM) Model.

• Merge them accordingly and keep the most relevant data for discovering relationships

• Fill the missing values, by guessing or ignoring.

Page 5: Mini project1 team5

The best place for students to learn Applied Engineering 5 http://www.insofe.edu.in

Process [contd.]

• Visualize the processed data to bring better business insights

• Test the data insights accordingly

LOYAL CUSTOMER Predicted Positive Predicted Negative

Actual Positive TP FN

Actual Negative FP TN

Predicting a loyal customer as non Loyal Customer (False Negative)Error metric: Sensitivity/Recall.

TP: True Positive TN: True NegativeFP: False Positive FN: False Negative

Error

Page 6: Mini project1 team5

The best place for students to learn Applied Engineering 6 http://www.insofe.edu.in

Derived Attributes

Recent Customer+

Frequent Customer+

Monetary Value---------------------------Most likely customer

Derived Attributes

FrequencyLF Transactions From the Website2

FrequencyAppTransactions From the Other application/other websites

FreqGamePlay Number of games played

RecencyApp Last time purchased from the application

RecencyLF Last time purchased from the LF

Recencydown Last time downloaded

Recency

Frequency

minRecencyCum360Min number of days Last time purchased within 180 days

NumHouseChildren Number of children per customer

TotalTimeGamePlay Total gaming time

Monetary Value

Page 7: Mini project1 team5

The best place for students to learn Applied Engineering 7 http://www.insofe.edu.in

Graphs

Houses with children between the age group of 5 and 10, across all game types, generate maximum Revenue. Target these age group houses.

Page 8: Mini project1 team5

The best place for students to learn Applied Engineering 8 http://www.insofe.edu.in

Graphs [contd.]

FavoriteSource

app cart lf

0M

5M

10M

15M

20M

25M

30M

TotalRevenueGenerated

Revenue VS FavSource

Sum of TotalRevenueGenerated for each FavoriteSource.

US customers seems to be the ones who buy/play more games. Hence, generate much of the Revenue compared to UK. 

Apps are the most favorite source of medium used to play games. Thus, maximum Revenue is generated through app.

Page 9: Mini project1 team5

The best place for students to learn Applied Engineering 9 http://www.insofe.edu.in

Graphs [contd.]

House with 2 – 4 children, irrespective of their age, seem to play more games. Hence, Families with 2-4 children have to be concentrated up on, as they are the ones who are generating highest revenue.

Page 10: Mini project1 team5

The best place for students to learn Applied Engineering 10 http://www.insofe.edu.in

Graphs [contd.]

Number of gamers who play games are gender unbiased. Frequency of game play distributions are similar for female and male children. Gender need be major point of focus.

Page 11: Mini project1 team5

The best place for students to learn Applied Engineering 11 http://www.insofe.edu.in

Graphs [contd.]

Most revenue is generated through customers who use credit card followed by USL.F com Physical Card and US Retail physical card. These customers can be considered for providing offers.

FavoriteChannel

AU/NZ Retail Physical CardBonus code for LeapPad/Ex.. Bonus Code for TRU LeapPa.. Bonus Code for TRU Value C..

Bonus for TRU1 Bonus for TRU2 Bonus for TRU3

Canada Retail Physical Card

Cheerios Bonus Code

Credit Card CS for TRU

Customer Service

Early Bird Bonus Code

Ireland Retail Physical CardLeapfrog.com Online Purcha.. LeapSchool Cart Bonus Code Marketing Promotion - The .. NON SHIPPABLE, LEARNING.. NON SHIPPABLE, LEARNING.. Rest of World Retail Physical..

Retail Hardware

Retail Leaplet CardsTarget Exclusive 2010

TRU Exclusive Book App Card TRU Exclusive Creativity Card

TRU Exclusive Game CardU.S. Bundle $20 Value Card

U.S. LF.com Physical CardU.S. LF.com Virtual Code (e..

U.S. Retail Physical Card UK Retail Physical Card

US Bundle (in-package) - Ta..

0K

50K

100K

150K

200K

TotalRevenueGenerated

Revenue VS FavChannel

TotalRevenueGenerated for each FavoriteChannel.

Page 12: Mini project1 team5

The best place for students to learn Applied Engineering 12 http://www.insofe.edu.in

Hypothesis TestNull Hypothesis: Average number of children in a house is 6 with 99% accuracy

Alternate Hypothesis: Average number of children in a house can never be 6 with 99% accuracy

Statistic: Z-Statistic, One attribute with more than 30 records

From the above graph, we are 99.99% confident that, average number of children per house can never be 6. Null Hypothesis assumtion is incorrect

Page 13: Mini project1 team5

The best place for students to learn Applied Engineering 13 http://www.insofe.edu.in

Code snippets• Packages: sqldf, DMwR, reshape

• Filter Data toDate30 <- as.Date('2012-01-26',format='%Y-%m-%d')

widCart_30 <- subset(stg_bgdt_cust_purc_app_updated,

stg_bgdt_cust_purc_app_updated$TRANSACTION_DATE < toDate)

• Missing Values Central Imputation used for filling central tendency values (KNN Imputation is not used as there are 108 dimensions with 1 lakh plus records, we tried it by running 2 days, but no use!!)

• Derive Data maxValueInRow <- data.frame(castedChannel$CONTACT_WID, apply(castedChannel[, -1], 1, which.max), colnames(castedChannel[, -

1]) [apply(castedChannel[, -1], 1, which.max)], rowMeans(withoutWID, na.rm=T))

#computes column name with with max value and maximum value plus mean value for eacsh row in castedChannel removing NA value

. . .

Page 14: Mini project1 team5

The best place for students to learn Applied Engineering 14 http://www.insofe.edu.in

International School of Engineering

Plot 63/A, 1st Floor, Road # 13, Film Nagar, Jubilee Hills, Hyderabad - 500 033 

For Individuals: +91-9502334561/63 or 040-65743991

For Corporates: +91-9618483483

Web: http://www.insofe.edu.in

Facebook: https://www.facebook.com/insofe

Twitter: https://twitter.com/Insofeedu

YouTube: http://www.youtube.com/InsofeVideos

SlideShare: http://www.slideshare.net/INSOFE

LinkedIn: http://www.linkedin.com/company/international-school-of-engineering

This presentation may contain references to findings of various reports available in the public domain. INSOFE makes no representation as to their accuracy or that the organization

subscribes to those findings.