something's a bit off…

@calmyourtities lol sleepy and daydreaming me forgot what the project is about

Ooo this looks fun and creepy at the same time lol

@turturtles I think it is way more creepy if they don't tell you but sell the information. Basically what every social media does. I think this project, if it comes to life, will be great to show people both the risk and the opportunity with AI data analysing.

@turturtles @jobylie It's pretty much an idea to tell you be careful what you post. This is what advertisers and other people you don't know can learn about your personal life

I'll try to contribute if I'll have some free time. Sounds very interesting

@dontPanic right now I should probably learn more how some stuff work with Instagram and start designing an interface and work on the computer vision part

@UnknownDev that's fine

I'll contribute to it, when I have free time again, as well. Never did such a thing before. But hey, you always learn something new.

We could probably start with a basic cli app and go from there? Use the API to pull in the images and user likes as a start?

@turturtles good idea. Simple then advanced

@turturtles that's a great idea. I haven't even thought about the likes

@turturtles we could also probably guess who their best friend is by either seeing who is in most of their photos, likes each other posts and follows each other

@UnknownDev haha I was just thinking we might need that data to get a whole picture, and was also just browsing the IG API documentation.

You can have a look at thr geo tracker and match the data with, for instance, google maps. Find out about their habits as well as interests. Unfortunately, you don't have all the data at your hand that ig or Facebook have: by finding out with whom you spend a lot of time (ofc also with whom you're friends with) they assume common interests. When somebody really likes cars but doesn't post a lot of photos of them, but spends a lot of time with people who like car sites and post photos, it can be assumed.

You could use the Google Vision API as developing your own neural network and training it with what objects are what might take a long while and Google has already done it. I believe they use the same technology with Google Photos to recognise objects within images. From there you can maybe go ahead and start creating some sort of a model on "learning".

Awesome idea, have you thought about hashtag analysis? I’d love to contribute!

@sticreations 🤔 I haven't actually.

🏕️

The project sounds really great but can I know what's the ultimate goal of this?
After the classifier identifies the emotions of a user where are you going to use this information?
Is there a bigger picture to this?
Very keen to know!!