# What kind of statistical analyses can I do with my data?

I'm trying to analyze human intentions in clicking google ad word keywords.

In this dataset I have the usual adword details, for example

CTR = Clicks / Impressions
CPC = Cost / Clicks
CPA = Cost / Converted Clicks
ROI = Total Conversion Value / Cost
CVR = Converted Clicks / Clicks
Weighted Avg Pos = WAP / Impressions


Basically, I can imagine a relationship between Keywords and CTR. I can do some psychological analysis, like opinion mining and emotion detection, like a vice. Still, I need to show different types of analysis.

So what kind of advanced statistical analysis methods can be applied to this data set? My data looks like this:

Keyword |Average Position| Average CPC| Clicks| CTR| Cost| Impressions

android app developers ,1 ,0.5, 21.79, 8.7%, 10.99, 250

This is what your post looks like to me:

"I have these data, what can I do with them?"

The answer is "A LOT". What do you want to do exactly? what is the goal?

From the topic, I assume your goal is to find the relations between "Keyword", "Average Position", "Clicks", "Cost", ... with "Impressions". If this is true, then you have multiple predictor variables and a target variable (Impressions) in which you would like to know the relation. This can be done through multiple statistical techniques such as:

1. Pearson Correlation
2. Spearman Correlation
3. Kendall Correlation
4. Mutual Information
5. Simple regression (and checking importance of each predictor variable)
6. RReliefF algorithm etc.

You have to wonder what is what you want to improve in your page. In my opinion, the most important analysis in every page is the conversion, this is people that enter the page versus people that buys something (or interact). Everything in the page should be optimized to increase this value.

Another important thing would be cost versus conversion. To see how the cost affects your conversion.

If your content can be share it is important also number of visits versus number of shares.