Time-series for website traffice analysis aftere adding influencer? On my website (interactivia.ro) I've added a gamification module from CaptainUp. I'm interested to find out how this gamification module influenced my website traffic.
The data extracted from Google Analytics is for an year and for this period I also have (almost) daily interactions from the gamification module, something like this:
https://goo.gl/fFZCJ1
My questions:
- how can I find out the influence (if any) of the gamification module on my website traffic
- is there a step-by-step example (in SPSS)?
Thank you,
Catalin
 A: It really depends on what kind of an effect you are looking for, but if you are looking for an increase or a decrease, then I would just plot traffic data as two separate time-series in Excel.  If you see a response, and want to understand if that change is statistically significant, then you will need to combine your data into before and after datasets (e.g., unique visits per day before and after), and then look for a shift in the mean via an ANOVA, t-test, or similar.  All of this should be doable in Excel or any stats package.
A: I took your 450 values (Y and X ) into AUTOBOX , a piece of software uniquely designed for your problem. In summary , you had significant error variance change requiring Generalized Least Squares VIA the Tsay Test  . This lead to the detection of a number of level shifts , significant days-of-the week effects ( and three changes in the daay-of-the-week factors ) and an ARIMA component reflecting period to period and week to week auto-dependency. . The residuals from this model are reasonable  and exhibit no autocorrelation . The Actual and Forecast plot is here with the forecasts shown here  . A table of forecasts is here  . In summary your X variable is important contemporaneously. 
