Yeah we don't get many likes...
What it does
Our program takes in some instagram information about the user, such as number of following, and a picture the user would like to post. The program will return the number of likes this picture will get instantly!
How we built it
We gathered more than 8000 posts from instagram and preprocessed the post image with Clarifai API. Then we trained an artificial neural network to see how many likes each picture will get.
Challenges we ran into
Data preprocessing is very time-consuming. We also ran into problems such feature scaling, limit amount of calls for Clarifai, etc, but in the end we were able to solve them and finish the project.
Accomplishments that we're proud of
We didn't have Keras set up on our desktop, but we were able to set up all API and get them to start working on time. Both of us have little experience of Machine Learning, and implementing the project from beginning to end is really a good experience. We hope to build more exciting projects like this.
What we learned
We learned the importance of teamwork and collaboration. This would not be finished if we didn't split the work well. We also get more experience on Deep Learning in general.
What's next for Instafamous
Our original idea is to make this an app, so that user can take a picture anytime and know the number of likes he/she might get instantly. If we have more time, this would be a great addition to our project and much more accessible.