Using Statistics to analyze news articles


I want to analyze 5000+ news articles by scanning all articles for specific words.
My dataset is going to look like this.
Articles - Page views - Bounce ratio - Session time - word 1 - word 2 - etc .
Article 1 - 4000 - 0,88 - 1,12 - 0 - 1 -

My questions are:
Which software should I use?
What kind of analyze can I use to tell me something on the best scoring words and word combinations in this articles?

At the end I want to predict what kind of news articles will score best online.

If you can point me in the right direction, it would be a great help.