Our project takes in an Instagram profile and uses Google Vision APIs to identify many different objects and landscapes throughout all of their pictures. From here we determine their biggest interests along with many of the locations the user has visited.
We were inspired by the fact that people put so much information on social media about themselves without realizing how much data is actually available. This was made in the sake of safety to show people how much information could actually be extracted in a matter of minutes. It could also be used to perform dictionary attacks on passwords.
It downloads images from a user's Instagram, and uses Google Cloud APIs to get keywords and common objects/scenarios on that Instagram. It then uses Google Excel API to generate a TreeMap of the common words after they have been parsed with pandas from JSON into csv.
Organizing all the JSON dating and getting it into a simple CSV format took a lot of effort because the Google API spits out a lot of different JSON files, and getting one combined word frequency is very difficult to manage.
We are most proud of using the Google Vision API to integrate machine learning easily into our project, and then using other Google APIs to generate interest visual representations of this data.
We learned how to use Google APIs, and pandas to parse all the JSON files. Both of these are very useful in the fields of machine learning and data science, something my teammate and I are both looking to get into. We learned about geolocation and Google Maps API as well.
Adding multiple features to show locations the user has been to such as an interactive map with pins dropped, and adding dictionary attacks to securely test a users passwords.
NodeJS, Google Vision API, Python