Announcing the Winners of Invite Only Data Challenge: OCEAN Twitter Sentiment pt. 2
We are excited to announce the winners of the first-ever invite-only data challenge hosted by Ocean Protocol! We received great feedback when tasked our data science community with the original sentiment analysis of the OCEAN token challenge, and now are able to share results from the second leg of this frontier.
To extend the initial excitement, we invited the top 6 submissions to return for an invite-only data challenge to delve deeper into the relationship between social media engagement and the price of the OCEAN token. This blog will detail findings from the 6-person, invite-only data challenge.
The goals of this challenge were to dive into data to understand the correlation and causation of the price movement of a particular token based on social media sentiment, engagement, and investor behavior.
The winners are as follows:
First Place — Mohammad Jamal ($2500)
- Mohammed pulled statistics from major social media platforms and ran them through correlation calculations and tables to find the best-performing posts, profiles, and users that contribute to media engagement regarding the protocol. One suggestion that stood out through Mohammed’s report was stated that in order to achieve a minimum 10 percent impact on the OCEAN price through social media content, creators should focus on generating content that attracts a larger number of comments.
- For the transaction side, Mohammed took 100,000 transactions from 25,000 different token holders and conducted analysis on transaction times, volume, unique wallets, and relation to the length of the given trend at that moment (bearish or bullish). By this, Mohammed was able to conclude there is a lag in the market reaction to users' sentiments and proceeded to build a machine-learning model that reflected the outcomes of each of the mentioned parameters.
Second Place — Matin Nahvi ($1500)
- Matin broke down public data from Twitter, Github, On chain activity, and Medium blog posts to gather data to be used for this second part analysis. Their data was neatly bundled into daily aggregates, then compared against the daily OCEAN token price shifts. Their choice of navigation tool was The Bayesian Rule List — a model with rule-based predictions for a clear understanding of our path. Matin split the journey, dedicating the initial 80% to training (2019–12–30 to 2022–03–28) and the final 20% to evaluation (2022–03–30 to 2022–10–21).
- Matin found more power leveraging Google Trends results to find a correlation between the OCEAN token price and Google trends. This was something that hadn’t been considered in the specifics of this challenge, that we found to be very valuable.
Third Place — Luca Ordronneau ($1000)
- Luca took a fundamental approach to sourcing the data by breaking it down into key factors: volume, sentiment, and user influence. The machine learning model made in Luca’s report took these key factors and features into account and ran them through a Tabular Gradient Boosting Model (XGBoost). This deployed hyperparameters tuning and cross-validation to ensure an effective and generalizable model. Through data regression, analysis, and machine learning execution, Luca was able to formulate an Impact Sentiment Score, touching on recommendations for investment behavior, market trend forecasting, risk evaluation, and strategic planning.
We were very pleased to see how invitees identified relevant factors, sourced the appropriate data, conducted a thorough analysis, and presented their findings in a comprehensive report. Their exceptional skills and expertise were tested and compensated for in this exclusive data challenge.
Using Data for Insights
In this competition, the task was to analyze the relationship between social media engagement and the price of the OCEAN token. By exploring additional factors and leveraging machine learning techniques, uncovered valuable insights into the complex dynamics of the cryptocurrency market.
A tool to quantify and predict volatility in capital markets based on social media sentiment is a long-term outcome of this initiative. Starting with the OCEAN token, for example, this forefront of sentiment analysis is set to move to a broader spectrum of crypto assets, tokens, and commodities.
Both live and expired Ocean Data Challenges can be found here: https://desights.ai/o/0x8264CE39F7532871868053198fCDC2E2C81d074a
What was the task?
Participants invited to the second leg of the OCEAN token sentiment challenge competed for a $5000 prize pool, and were tasked with the following:
- Identifying three key factors that could potentially influence the price of the OCEAN token, considering social media engagement.
- Provide a detailed description of the data sources you would consider for each identified factor. Be specific about the datasets, APIs, or platforms you would utilize to obtain the necessary data for your analysis.
- Design a methodology to measure and analyze the relationship between social media engagement and the price of the OCEAN token. Explain how you would collect and preprocess data to conduct sentiment analysis on platforms like Twitter or any other relevant social media platform.
- In addition to social media sentiment, outline their approach for exploring the impact of the external factors you identified in question 1. Describe necessary data transformations, calculations, or statistical techniques you would employ to analyze the relationships between these factors and the OCEAN token price.
- Develop a machine learning model that can provide insights into the relationship between social media engagement and the price of the OCEAN token.
- Describe the ML model you chose and explain why it suited this task. Outline the steps involved in training, evaluating, and interpreting the model’s predictions. Include details such as the choice of algorithms, feature engineering techniques, model training methodology, and any considerations for handling potential challenges, such as data imbalance or overfitting. Explain how the ML model contributed to your analysis and supported your findings in the report.
This was the first token sentiment analysis and invitation-only challenge hosted by Ocean Protocol. Future data-driven social media data challenges can be expected for other prominent Crypto and DeFi Protocols.
Do you think your project could benefit from social media sentiment analysis? We encourage you to explore our community of data scientists and crypto enthusiasts and examine the benefits of running a data challenge on your protocol with Ocean Protocol Data Challenges.
About Ocean Protocol
Ocean was founded to level the playing field for AI and data. Ocean tools enable people to privately & securely publish, exchange, and consume data.
Follow Ocean on Twitter or Telegram to keep up to date. Chat directly with the Ocean community on Discord. Or, track Ocean progress directly on GitHub.
