It seems that you're using an outdated browser. Some things may not work as they should (or don't work at all).
We suggest you upgrade newer and better browser like: Chrome, Firefox, Internet Explorer or Opera

×
Community
GOG GALAXY 2.0 discussionSorting games in a machine learning way(3 posts)(3 posts)
(3 posts)
Pages:
1
This is my favourite topic
Posted September 03, 2019
First of all, a question... are the info about games stored locally in anyway? If so, can a user have access to their own? Or do you have APIs to pull the info from the cloud (and from other platforms) on the fly?

I ask because I have an idea do develop my own particular search. You see, I have around 500+ games in total. I usually choose the one to play based on how I feel at the time, time of the day, etc... but it would be nice to sort the games in a machine learning way, as an added way to suggest games for me to play (think netflix or even steam suggestions).

BUT for that I need to know if and which info I'd have access to to make a reasonable ML model. Being a data scientist myself, I have some ideas already. Even if I don't have easy access to the games info, I'd probably still be able to 'scrape' around as needed.

But just to be clear: I'd only develop that for myself, using my own "data", so there'd no legal problems. Unless, if I'm successful of course, the gog galaxy dev team wants to collaborate. :)

Cheers!
Posted September 04, 2019
avatar
GenlyAi: First of all, a question... are the info about games stored locally in anyway? If so, can a user have access to their own? Or do you have APIs to pull the info from the cloud (and from other platforms) on the fly?

I ask because I have an idea do develop my own particular search. You see, I have around 500+ games in total. I usually choose the one to play based on how I feel at the time, time of the day, etc... but it would be nice to sort the games in a machine learning way, as an added way to suggest games for me to play (think netflix or even steam suggestions).

BUT for that I need to know if and which info I'd have access to to make a reasonable ML model. Being a data scientist myself, I have some ideas already. Even if I don't have easy access to the games info, I'd probably still be able to 'scrape' around as needed.

But just to be clear: I'd only develop that for myself, using my own "data", so there'd no legal problems. Unless, if I'm successful of course, the gog galaxy dev team wants to collaborate. :)

Cheers!
Since GOG pulls data from IGDB (it's not known whether only temporarily for the purposes of beta testing or permanently)
You can experiment with IGDB database API . https://www.igdb.com/api
You may also want to share your ideas with IGDB developers at their discord channel. https://discord.gg/WvBNFRu
Posted September 05, 2019

Since GOG pulls data from IGDB (it's not known whether only temporarily for the purposes of beta testing or permanently)
You can experiment with IGDB database API . https://www.igdb.com/api
You may also want to share your ideas with IGDB developers at their discord channel. https://discord.gg/WvBNFRu
Perfect, thank you. I'd then need to match with my own list of games for such queries and perhaps find out info related specific for the "user" (me) for each game, like total hours played, achievement of the type 'completed game', etc. As a last resort, I could probably just crawl and parse my own game web pages for that.

Anyway, thanks a lot! If and once I create such a project, I'll set up a github if more people get interested.

Cheers
Pages:
1
This is my favourite topic
Community
GOG GALAXY 2.0 discussionSorting games in a machine learning way(3 posts)(3 posts)
(3 posts)