Believing in a Number — Exploring conversations that impacted a decentralised cryptocurrency.
Experience Design
Data Visualisation
Front-End Development
School Project
In this school project, I experiment with different methods of showcasing different types of data points (numerical, textual, etc), allowing data to be communicated in alternate ways for social influence.
To do that, I visualise how the emotions of comments about bitcoin affects people’s decision to buy and sell bitcoin, in a three-part interactive web experience.
Team
Yuan Jie, Designer
Under the supervision of
Dr Clement Zheng.
What I did
Design Research
Data Cleaning & Analysis
User Testing
Prototyping
Results
Demonstrated the working prototype, visualised over 100 resulting game data points.
Practised MERN stack for full-stack development.
Sparked various conversations about cryptocurrency during demonstration.
Design Process
1. Discovery
Research on interesting and open-source datasets.
Ended up using a dataset of 500000 comments on a social news website, Reddit. Find it online here.
2. Dataset Analysis & Clean Up
Uncovered the sentiments behind each comment using the NRC-EmoLex library.
Visualised results for trend spotting.
3. Prototyping & User Testing
Meaningful data and trends were visualised in experience prototypes.
Several rounds of user tests were done to validate hypotheses.
3. Final Prototype
Final prototype was demonstrated to players.
Discovery
Cryptocurrencies, just numbers that matter?
Whilst searching for datasets, I stumbled upon something I've dabbled with for awhile, cryptocurrencies.
"You're visualising a data set about something which in itself is made of data!" was something interesting I remembered my professor said.
That sentence really stuck to me (as you can tell, I'm even writing this 1 year later).
Observation
There was an interesting duality, between how cryptocurrencies (essentially made up of numerical data) are worth so much to people.
Moreover, being decentralised it really showcases how much people bought into this idea of a vision.
Data Analysis
Finding what these numerical fluctuations tell us.
Unlike traditional currencies, Bitcoin’s value is not backed by a tangible object or an organisation. Its confidence comes in its perceived value. This makes Bitcoin prone to high volatility in its price action, solely depending on public opinion. Can these opinion tell us how confident people were in Bitcoin?
Using Sentiment Analysis*, I was able to uncover the emotions behind each comment.
By putting the result against the price of BTC/USD, I wanted to explored how much of an impact emotions made in the price action of Bitcoin during 2017 to 2019.
Sentiment Analysis* — The use of natural language processing, text analysis, computational linguistics, and biometrics to systematically identify, extract, quantify, and study affective states and subjective information.
Further Observations
How do people react to these fluctuations?
To find out if the price action really do play a role in logical decisions, I started off with a mini experience prototype with a group of 8 classmates.
They were asked to sit through the history of Bitcoin's price throughout 2012 to 2019, riding through the volatility of the price action while showing relavant "positive" and "negative" comments.
This prototype showcases scrolling comments sorted against positive and negative sentiments. It also shows how the price of Bitcoin changes while comments of negative and positive sentiments appear.
How did people feel when using the prototype?
Insight 1
The price action of cryptocurrency was the most intriguing to most people because it was not commonly seen.
People who knew about cryptocurrency were consistently intrigued about Bitcoin's price action, while people who did not have prior knowledge of Bitcoin found it hard to understand, but were also soon intrigued by the hype around its insanely volatile numbers. This is because of how huge the price action were.
Insight 2
Textual data take time to be digested and understood.
Users did not interact much with the website as they thought it was “auto scrolled”. Thereafter, users who tried to read the comments were too slow to catch up to the scrolling speed. Hence, they did not understand what the data meant.
Reiteration
Will gamification enhance importance on price fluctuations?
The second prototype introduces a game mechanic to the experience, where players are challenged to buy and sell Bitcoin, while exposed to comments surrounding cryptocurrency. Players are able to view their score at the end to spark conversations around the volatility of cryptocurrency.
A voice synthesiser announced each player’s loss and gain to everyone, to add social interactivity in the game.
Insight 1
The high amount of interactivity in different mediums (text, speech, play), the lesser chance of digesting textual information because of overwhelming cognitive load.
During the game, players were more focused on playing the game than reading the comments due to distracting audio. Players discussed their profit afterwards, but did not discuss about cryptocurrency. Basically it became a meme game where everyone was spamming “buy” and “sell”. :(
Understand the Method
Understand complex nature of sentiments in texts.
Firstly, I allowed my users to experience sentiment analysis through social interactions. To do that, I designed a live experience where people can quantify their own feelings. In a chatroom, users are able to send how they are feeling to everyone else.
These comments are posted live, mimicking the open community of Reddit. Sentences with a higher positivity score are given a higher visual hierarchy, allowing users to be immersed by a sea of positive comments.
Understand the Context
Reading and understanding those Reddit comments.
Secondly, I designed a live search bar where users are able to search for interesting comments to their liking to allow the opportunity for more meaningful comments to surface, using the act of sorting the dataset as an interaction.
The act of finding comments affords a deeper understanding of my dataset and is designed to influence their buying decision of Bitcoin.
Applying the Knowledge
Simple gamification to show the impact on price and emotions.
Lastly, I designed a mini game where users are challenged to buy and sell Bitcoin, according to the sentiment score of each comment.
By reimagining trading without the price, I was able to showcase the scale of the influence that sentiments have on trading Bitcoin. At the end of a round, profits that they gained are revealed, and users can then compare them to other player’s profit.
What i've learnt throughout this semester project.
Lesson
Communicating trends and textual data can be powerful because it can provoke meaningful discussions, especially when the data point originates from people themselves.
What made my dataset special was that textual data included sentiments showed what people actually felt on that particular day and date about cryptocurrency. This relatable data can often pique curiosity, "Do what we say really matter?".
Lesson
We should always keep the level of cognitive load in mind when designing experiences because the more cluttered and fast-paced an experience is, the less time someone would have to pick (important) information up.
Gameplay is an interesting way of engaging stakeholders. It is important to juggle the playable experience, where players might be more focused on playing the game than reading the comments due to distractions.
At one point, my prototype basically became a meme game where everyone was spamming “buy” and “sell”.
