The What’s Good Project

The information we consume shapes our perception of reality.

K. Alexander Ashe

I had a conversation with a customer about the plethora of negative and fake news. The customer asked me if the Rossella algorithm could help them find good news. I thought this was a problem worth solving because Rossella was founded to combat bias and false perception proliferated by bots, trolls, and fake news. Challenge accepted!

I recently attended a Google Cloud OnBoard training session in DC. I starting thinking about Google Cloud’s Machine Learning capabilities which enable analysis, insights, and pattern recognition for data, text, video, image, and language translation. Google has a Natural Language Processing API. I dug into the technology and learned that Google ML NLP text analysis includes entity recognition (person, place, thing) and sentiment (negative to positive) analysis.

The sentiment analysis analyzes text and calculates a score from -1 to +1 as well as a magnitude. The score indicates whether the overall tone of the text is positive or negative. Adding sentiment and entity analysis can be combined to determine the tone of text with regard to specific persons or organizations. I did find the Python implementation to be a lot more efficient than PHP — my go-to coding language.

Google’s Sentiment Analysis and the Rossella algorithm can enable people to share recent good news from trusted sources. The information we consume shapes our perception of reality. For instance, according to FBI statistics, violent crime is less than 14% of all reported crime but it makes up between 50–80% of the local news coverage according to the Dart Center for Journalism and Trauma.

Inundating oneself with more good news is not just an emotional necessity-it’s is also important so that we can re-frame our perception of reality. Rossella launched the What’s Good project to encourage people share more good news.

Rossella Machine Learning — Tweet Your Topic and Tone

The What’s Good Project