Panel data analysis - Granger casuality

Dear all,

I'm doing research about relationship between energy consumption and GDP growth of European countries in timespan of 20years.
I would like to test a hypothesis : energy consumption does affects GDP growth. (if does, how? Unidriectional, bidirectional)
As I searched through articles on this topic, most of them were using method of Granger casulity, so I decided to take this method, too.
Firstly I should make some test for unit root (to check if my variables are stationary or not)
Afterwards, some cointegration test should be applied (Pedroni, VAR, VECM, ARDL)?
Finally Granger casuality test should be runned.
Here is the link to article with similar procedure :

Does anyone have experience doing this kind of analysis?
Is it better to use STATA or Eviews?

Should I use panel data? I would like to have my results seperated by each country.

Thanks a lot!