Remember the time when you wanted to look up that picture of yours from the beach 5 years ago and couldn't think of the folder, forget the name of that image - DSC_0808.jpg (Ewww!)? Sounds familiar? We all have a whole lot of pictures, from a few hundreds to even several thousands, making it difficult to track down a particular one. Pictures and videos are often associated with the memories of the places that they were taken at and if there was a tool that would help look them up easily, it would be dope!

What it does

Fotolitic imports pictures from your Facebook account and builds and generates a tag repository from your images, making them searchable without you having to tag them manually. It also helps you get an idea of the relative number of tags in your account by presenting some neat data visualizations. It helps you browse your gallery like never before.

How we built it

Technology stack - Python(Flask), D3.js and Elasticsearch. We used the Facebook OAuth API to authenticate a user and generate permissions to access the user's gallery. We retrieved images, used Clarifai's API to fetch tags from these images and indexed them using Elasticsearch. From this data, along with the geolocation metadata, we were able to generate some data visualizations using Esri's API as well as D3.js.

Challenges we ran into

We ran into some integration issues, the JavaScript code broke several times, forcing us to move some of our code to Python to complete the pipeline.

Accomplishments that we're proud of

We are proud of the reception of the idea and are very happy with the progress that we have made.

What we learned

We had a good time building our Flask and Elasticsearch skills.

What's next for fotolitic

Connect with other storage providers like Google Drive, Dropbox, etc. Work on the scalability and stability aspects of the application and of course release a beta version soon!

Share this project: