Extract useful patent numbers for beginners

[no_idea]

Hey PATSTAT-community,

I'm currently thinking about starting my thesis to the topic: influence of a firm's innovation activities to it's market value.
I thought about how to measure innovation, and thus, among other things, I'm looking for patent data.
Now I found my way to this database.

Unfortunately, I have no experience with setting up datasets, extracting data and definitly need help.

So can I extract patents to specific firms and years from this database? ("count patents for firm i in year t"?)
What about trademarks?
How difficult will it be to get useful data?
What skills are required? What kind of software/tool?
Do my little experiences in "MySQL"-queries suffice? Or are they even relevant?
How much time should be scheduled for this?

So you see, I have no idea how i should approach this. I would really appreciate your support.

[Geert Boedt]

Hello,
A good start is to go through the various documents (and self learning modules) that are available on the PATSTAT webpage.

Extracting a data set that gives you "number of patents per company" is rather straightforward. Although most researchers would use a count of "families" in order to avoid duplicates due to similar filings for the same invention in another country. (or you could also count "first filings" if the country where the patent was filed is not relevant.)

Trademarks are not available in PATSTAT. Some researchers have linked EUIPO trademark data to patents, I would recommend you to look up those studies and contact the researchers directly. They can maybe share their data or methodology.
On skill and tools: if you use PATSTAT Online, you will need SQL skills; depending on the complexity of the data you need, the queries will become more extensive upto a level that you an not use PATSTAT Online anymore. In which case purchasing the PATSTAT data and installing it on a local server will be the only solution. That will give you full freedom to work with the data using whatever tools your are familiar with. (You should discuss this with your the IT unit of your university, and see if they can host it for you...)
Time: impossible to know without more details. Bottom line is that if you make mistakes in the first data aggregation steps, it will have a knock-on effect for all further analysis. Ideally, you should also have a basic knowledge on the patent grating procedure.
A simple graph or data set is not much work. (Or you can maybe use the figures that we publish on the EPO webside ?) More complex analysis will require more data aggregation, maybe some cleaning procedures before you have a good starting set of data you will use for your research.
Maybe you can give some more details ?