(Views + Shares)

We have added up all the shares & Views from columns, namely:
  1.  Twitter shares
  2.  Facebook shares
  3.  Vine shares
  4.  Custom shares
  5.  Email shares
  6.  Three second views
  7.  Auto play views
  8.  Video plays
  
to create a new column 'Views + Shares' for our analysis. This is a collective measure of a 
user's digital activity on the add.

Country wise click rate percentage (clicks/visible impressions) conditioned on Brand Guarantee (BG)

Observations:
    1. 'FR' (France) has the highest click rate across all geographies once Brand guarantee
    gets into action
    2. 'ES' witnessed the largest difference in click rate before (orange) & after(blue) 
    Brand guarantee

Country wise (Views + Shares/visible impressions) percentage conditioned on Brand Guarantee (BG)

Observations:
    1. Only North American countries ('US', 'CA') have a significant (Views + Shares) rate
    2. The trend is counter intutive in the above 2 countries, as (View + Shares) rate has decreased 
    under the presence of brand guarantee

Device Category wise Filled pages rate % (Filled Pages/ Total Impressions) by Post location

Observations:
    1. Filled Page rate is similar for all location types for any device
    2. Average Filled Page rate has been nearly 23% on the whole

Device Category wise (Views + Shares/ Visible Impressions) % by Post location

observations:
    1. 'Tablets' have the highest (Views + Shares) rate amongst the devices
    2. 'Other' has the highest (Views + Shares) rate amongst all locations

Dispersion of Mean Total Impressions requests and mean Impressions for each publisher device wise

***************

## .
## .
## .
## .
## .
## .
## .
## .
## .
## We witness a linear dependency between the mean of Total Impression requests and Impressions. So 
## let us quantify the correlation between these 2 quantities and the significance between the 2 quantities,
## using a simple Correlation test
## 
##  Pearson's product-moment correlation
## 
## data:  data.6$T_IMPRESSIONS and data.6$IMPRESSIONS
## t = 119.93, df = 4508, p-value < 2.2e-16
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  0.8654166 0.8793553
## sample estimates:
##       cor 
## 0.8725635
## The test proves, that there is a significant corrrelation between Total Impression requests & Impressions.
## Their correlation is .873.