You are on page 1of 4

Experimentation

Cause and effect analysis

Exp is a process of rejection or accepting a hyprthesis

One variable giving ride to other -causation

Co relation two variale behave I the same way

Rscript+

Cc<-read.csv(file.choos(),heder=True)

Check for NAs

Sum(is.na(cc)

Understand user division

Library(dplyr)

Cc %>%

Group_by(version) %>%

Summarise (user=n())

#commo metirx is retention

Understanding the retention performances ‘

Cc%>%

Group_by(veions) %>%

Summarize(ave_1=mean(retention_1),ave_7=mean(retention_7)

Retention for each AB group

AB<-cc%>%
Group_by(version)%>%

Summarie(users=n(), ave_1=mean(retention_1)

Ave_7=mean(retention_7))

Graph the data

Liberey(ggplot)

Ggplot(AB, ase(x=version, y=ave_1), geom_col(aes(fill=version)

Ggplot(AB, ase(x=version, y=ave_7), geom_col(aes(fill=version)

#we need to know is there a differen in retention

Testing Part.

#state the hypothesis

Null hypothesis= No difference in the retention of both version

Alt hy= there is a difference in the retention rate of

Step 2+++Determine the type of test

Is it a two or one sided test?

It is two sided since there is no directionality involved

# is this is sigle/independent sample test or double size sample

This is an independent sample because there two independent group involved

D=step++3

Determine the test Statistics

R1<- t.test(retention _1~ version ,data=CC, var.equal=TRUE, paired=FALSe)


R7<- t.test(retention _7~ version ,data=CC, var.equal=TRUE, paired=FALSe)

Step 4+++

Report the results

# retention on either gates of version after 1 day is as follows

45% and 44% but it is an insignificant result

#retention on either version after gate 7 is a follows 19%(g30) and 18%

And it is significant result

State your conclusion

So if you want to keep retention it is for level 30

So if want to keep retention not move from 30 t0 40

There is a significan dip if move from 30 ----40

Reject if CI is 90%

You might also like