Milestone 2 Submission for Orbital

Team Name:

Tomatinator-3000

Proposed Level of Achievement:

Apollo 11

Motivation In this era where fake news is rampant, we want to create an app - TweeterBuddy, that helps users check their tweets for fake news. Moreover, social media cancel culture is brutal and we want to help users avoid it by giving them an idea of their tweet’s sentiment, check for grammatical errors and offensive language. The last thing we want is for our users to lose a million followers overnight! Furthermore, we want the Internet and the social media space to be more inclusive, friendly and warm. This app will encourage users to use kinder language on the micro-blogging website Twitter and in turn make the Internet a safer space. Aim We hope to create an application to enhance the utility of twitter by giving people an insight on how their tweet looks before posting it. User Stories

1. As a journalist, I want to tell the correct information to the public.
2. As a ‘netizen’, I want to create a culture of acceptance and positivity online.
3. As a Twitter user, I want to ensure I don’t consume fake news or propaganda.
4. As an influencer, I want to ensure my tweets are not offensive to my followers. Scope of Project We are planning to make a mobile application which acts as an enhancement of the twitter app. Our application will have the basic UI which allows users to input a text and find out what the sentiment of the text is, and whether or not the text contains fake news. The “display history” button allows users to look at their previous inputs, see the distribution of words they use for each sentiment and type of news. The “display history” option will provide the user with multiple graphs, highlighting their most used hashtags, and a wordcloud with their most commonly used words for each sentiment and type of news. Currently, the app has a textbox where users can enter their text and get a fake news and sentiment analysis. The interface of the app and the history are under development.

Features and timeline The application holds a database containing previous tweets and their analysis. It makes the user more aware and considerate about their tweets, making twitter a safer and reliable website as a whole. Main Features: A textbox for the user to input their text in A display history feature to showcase the user’s last 5 positive, negative, neutral, real news and fake news tweets in a graphical form

Features to be completed by June (Milestone 1 & 2):

1. A textbox to write text
2. Sentiment analysis of the tweet
3. Fake news detection of the tweet Proof of concepts: examples of our app

Things to note We have yet to create a more sophisticated UI for the app. Currently the app is a very rudimentary version of itself. We plan on imitating Twitter’s UI and allowing the user to post tweets to twitter in real time from the app. The fake news model took longer than expected to be developed so UI could not be developed properly. The sentiment analysis model has been completed and a telegram bot has been made. Check out @TomatinatorBot on telegram t.me/TomatinatorBot Features to be completed by the end of July (Milestone 3): · Complete development of the UI, making the app more interactive and pleasing to look at · Integration of a real-time database, so all tweets made by the user can be stored · “Display History” feature, which will allow the user to analyse their tweeting habits

How the features will be implemented: For the frontend, we use flutter. The complete twitter UI can be remade in flutter. We chose flutter as it is flexible and can develop an app for both Android and iOS simultaneously. We also use python to create the models and the api for the models For the creation of the models, we use TensorFlow and tweepy, as well as WordCloud python libraries For the database of users’ tweets, we use Firebase as it is flexible, real-time and has fast in-place updates.

Tech Stack · Python · Flutter · Firebase (database to store users’ previous tweets)