Our inspiration for the project was to help fight the terrible problem that is suicide. A smart city is a city where the people's way of life is intelligent and where the people themselves are happy, and that is why we thought that fighting suicide was a way to make cities smarter. People tend to send cries for help on their social media, but they get so drowned by the massive amount of content that people tend not to notice till it's too late that someone was in need of support. That is something we wanted to amend.
What it does
Our program streams live tweets and bases them through a microsoft azure algorithm to evaluate if a single tweet is negative and indicative of sad thoughts or not. We store the users who wrote these tweets into a database, as well as the tweet themselves. If a user send too many red flags too our database, he is sent to our website to be evaluated by a professional to establish if the user is having suicidal thoughts or not. If the professional establishes that the user is at risk of causing harm to himself, he can click a button to send a reply to the latest tweet with a message of support and the link to a helpline. On the other hand, if the user does not seem at risk, the professional can delete him from the database.
How we built it
We built this program mainly using nodejs, multiple APIs such as Twitter's, a mongodb database and Microsoft Azure's text-analysis program.
Challenges we ran into
We ran into multiple challenges, mainly the acquirement of data to use to train our own model.
Accomplishments that we're proud of
We are proud to give in a finished product
What we learned
We learnt how to use different APIs and how to use Azure's services
What's next for Helpline Bot
We need to create a more accurate model and add additional cities so that we can strive to help more people.