EndenDragon / INFO201-Group-Project

The Dig on the Unofficial UW Discord Server

An INFO 201 Group Project, Spring 2018

Project Description

Dataset

We will be using the dataset from Discord App throughout the project. The Discord platform is built with gamers in mind, but because of the sleek user interface and their excellent freemium model which does not interrupt basic features of a chat software, many people have adopted Discord as their primary go-to place for their social interactions, even if their topic is completely irrelevant to games.

This project was ever-so-slightly inspired by the pisg software, the IRC Statistics Generator written in Perl programming language. Although the user interface of pisg is outdated along with the IRC technology itself, the statistics and graphs generated still holds to this day.

The data we will be working with is dynamically collected by Discord. We will access the data via the Application Programming Interface (API) offered by Discord. The data is not static; it is live and will change with activities of the users of the Discord platform.

Target Audience

The intended audience would be users whom are curious about the Discord channel of University of Washington (UW) in general. New members, for example, may be wondering how active the chat would be before taking a dive and joining in on the socialization. Current Discord users may find it intriguing to know whom is the most active user.

By the conclusion of the project, we would hope to answer these following questions:

• When is peak hour(s) of the given days?
• Who are the most active user(s) in the server?
• What are the most spoken words?

Technical Description

Data Retrieving

By using Discord API, we can retreive different "objects" that the Discord client keeps track of. We will primarily be interacting with the messages endpoint to read the messages objects since the last few months of the server's existence. The user endpoint also provides details about certain users, if needed. The messages endpoint provides timestamp of when the message has been sent. This can be useful for determining when the peak hour of the given day.

The documentation for Discord API can be found here: https://discordapp.com/developers

Data Wrangling

In order to keep the API calls to Discord to a minimum, our team would need to do an initial scraping of the messages and save it as .json files. Doing this process would prevent us from hitting the API rate limit in the long run.

To answer the questions we wish to solve for our audience, we will need to transform some of the data gathered from Discord API. For instance, in order to answer the question "what are the most spoken words", we will decompose every message we read from the API into separate words, count occurrences of each word, and store the statistics into a new data frame.

Libraries Used by This Project

• We are using tm, especially its removePunctuation() function, for removing punctuation marks from chat messages we scraped from Discord API.
• We are using pluralize to transform words in plural form into singular form.
• We are using shiny, the library for Shiny app, to present our product.

Anticipated Challenges

Some major challenges that we anticipate would be group coordination and wrangling with Discord's data. Building an awesome front end is also crucial to success in presentation. By using Shiny for our project, we can utilize its interactivity and let the user be in the driver's seat. Such interactivity like a custom date range would enhance the presentation.