extension in construction
[AS OF 21/03/2017 the public API has been suspended, to prevent abuse, due to lack of rate of access protection.]
Source code is openly available on: https://github.com/jgege/CrackTheMedia
Inspiration for the project was Facebook's recently developing feature of identifying fake news on the social network. We wanted to create something that would help people build tools that point in the general direction of media outlet transparency, and halt of fake news propagation.
What it does
The API provides three main tools, derived from google cloud services, with custom tools made to deduce data from the JSON's those google cloud service api's provide. The api provides hosted tools to determine the organization the outlet belongs to, the degree of originality of the images it used in its articles, and a specific article's accuracy, compared to other media outlet's output on the topic.
How we built it
We used php to make the requests to the google cloud service api's. We reduced the responses to data that we deemed useful for the implementation of the API in the given time frame. We used the yii framework to provide a quick base for our project and we hosted the whole thing on an Amazon Web Services es2 linux micro instance.
As a quick finish to the project, we built a very simple (and somewhat defective) chrome extension that uses the API we built.
Challenges we ran into
We used the google knowledge graph API, which we had some trouble making it behave the way we wanted it to.
Accomplishments that we're proud of
Two members of our team are complete beginners in the sphere of web development. For us understanding how things work was more important than anything else. I think we managed that on a fair level.
What we learned
We learned basic usage of some of the Google Cloud Services API's, and we also got insight of the direction that information providers like Google are taking.
What's next for Crack the Media API
We would love to make the API fully functional, without exceptions, and providing all the insight it possibly can. Also, while the image originality method works, it is not providing the data needed in a convenient format, and requires a lot of fiddling around, hence its exclusion from the demonstrator chrome extension.