We need to do analysis on various search terms and their Google Trends counts.
The script needs to be scaleable/configurable, which means that it gets the data to be searched for from a .csv configuration file that enables the following input:
Country (where the Google searches originate) / Search Term(s)
The country designates the countries from which the searches have been originated.
For each country, we want to make a separate Google Trends analysis of how search terms have been polled from the users over time. The idea is to get maximum available history for each Country / Search Term combination and write the numbers for each of these in a separate CSV file, with daily or weekly granularity.
Important: the search terms given below in English need to be translated to each countries dominant local language.
Which means the output for each row in each file consists of the following:
a) the date in the format MM/DD/YYYY and then
b) the value of the Google Trends count for that day for the search term (for that country). The column separator shall be a semicolon.
The script shall run daily. If the file is not yet existant, it shall write the entire data series from the start of its history to today. If there is an already existing file, it should only overwrite / add the new values for the last five days to the existing file.
The countries from which searches are to originate are the following:
China (but is there meaningful Google search activity in China, or is this all on Baidu?)
The various search terms are several dozen per country and they need to be localized fro each country (please use Google Translate).
These search Terms are in the following groups and concern the following terms:
Flatscreen OR television
"video game console"
travel OR vacation OR Vacations OR "Airline ticket" or "holiday trip"
"unemployment benefits" OR "social security" OR "employment exchange"
If there is an "OR" between search terms, that means that they should be searched with a boolean "or" logic.
Results: For each country of "search origin", we need to search for those 23 search terms, generating 23 seperate .csv files as an output.