-3

I have time series data for two year

12376 167827  3454596 9676112 342102 1232103 546102  5645696 96767110 23423119 
4577140  45435158 56767138 635435167 35443160 34534166 3213133 2132148 2342130 
7656127 43234117 56545130  5645138 56455149

And I want to identify the virality growth rate and then plotting the graphs for growth rate. Nothing is give like number of invitation sent and Percentage of Conversion to invitation. I wanted to know how I can achieve this and is there any package available in R for this.

Saurabh
  • 867
  • 3
  • 13
  • 28
  • 1
    -1 You have not given anything like enough information to go on, or made clear your question – alexwhan Mar 12 '13 at 08:14
  • @alexwhan what more Information you need please let me know. I have already mentioned that I have no information about how many invitation send and Conversion of invitation. I have given only time series data for 2 years. – Saurabh Mar 12 '13 at 08:17
  • 1
    @Samraan the issue here is that we don't have the context to understand what do you mean by "invitations", "Percentage of Conversation" or "virality growth rate". We don't even know what those numbers mean. – plannapus Mar 12 '13 at 08:19
  • @plannapus Virality is something like chain system means one user give invitation to say 3 and these 3 may gives invitation to 9 and so on. Percentage of Conversion means out of sent invitation how many have accept the invitation.finall Virality growth rate means at what rate user being added up. Hope I am clear now – Saurabh Mar 12 '13 at 08:24
  • 1
    Ok but what do the numbers in your time series mean? Are they a number of invitations, the number of invitations accepted, something else? – plannapus Mar 12 '13 at 08:30
  • The total number of customers (assuming you're talking about viral marketing) at a given time maybe? – plannapus Mar 12 '13 at 08:36
  • @plannapus yes data is about the total number of customers per months for 2 years. – Saurabh Mar 12 '13 at 09:31

1 Answers1

3

Although this question is very likely to get closed in the coming hours if you dont make your problem clearer, this might get you started at least:

mydf <- scan(textConnection("12376 167827  3454596 9676112 342102 1232103 546102  5645696 96767110 23423119  4577140  45435158 56767138 635435167 35443160 34534166 3213133 2132148 2342130 7656127 43234117 56545130  5645138 56455149"), )
plot(mydf, log="y", type="l")  # Gives you an overview of your time serie (with log axis)
gr <- diff(mydf)/mydf[-length(mydf)] # Gives you a growth rate between each of your values.
par(new=TRUE)
plot((1:(length(mydf)-1))+0.5, gr, type="l",   # Plots your growth rate
      col="red", axes=FALSE, xaxt="n", yaxt="n")
axis(4)
plannapus
  • 18,529
  • 4
  • 72
  • 94
  • Thank you for this answer, I think the main "trick" in this scenario is the diff(mydf)/mydf[-length(mydf)]. Really neat! – guerda Nov 26 '14 at 13:49