Data365 API for the Doctoral Research: How Can Tweets and Reddit Predict the Market Move?

It wasn’t just a dissertation, it was a dual-purpose mission.

• One part: publish meaningful academic work.
• The other: bring modern financial behavior into the classroom through live market data, case simulations, and better tools for his students.

Defining the Mission and Framing the Goal

To explore the connection between investor sentiment and market behavior, Professor Willis set out to:

[[ggp
https://cdn.prod.website-files.com/64d9f528c6b6445ddaf33590/6911f42d221d529c586a27df_Frame%201000003970.svg
https://cdn.prod.website-files.com/64d9f528c6b6445ddaf33590/6911f42dc5da0e641239b6cf_Frame%201000003971.svg
https://cdn.prod.website-files.com/64d9f528c6b6445ddaf33590/6911f42de36326bd37dedfdb_Frame%201000003972.svg
]]

• Analyze a full year (May 2024 – May 2025) of Reddit and Twitter posts mentioning main U.S. stocks ($TSLA, $AAPL, $AMZN, $NVDA, $GME, $JPM, $WMT);
• Filter those posts for financial sentiment terms like “bullish”, “bearish”, “buy”, “sell”;
• Compare social media sentiment scores (processed using VADER) against stock performance metrics (price movement, trading volume, and volatility) from sources like Yahoo Finance.

[[/ggp]]

All in the pursuit of two key questions:

1. Is there a statistically significant relationship (60%+) between social sentiment and market direction?
2. Does real-time social data improve the accuracy of short-term trading predictions?

The answers?‍
Yes, Roston T. Willis Jr. found them and made his project work.

The analysis confirmed a statistically significant link between social sentiment and market movements. And Data365 API was there to make it possible, providing the solid data foundation behind every insight. 

And here we are to tell the story to the end.

When Research Plans Are Ready but Data Falls Short

The ideas were there. The research questions were clearly defined. The analytical models were ready.
The data? Not so much. 

The data was the limiting factor and the main bottleneck for Professor Roston T. Willis Jr.

Professor Willis tried public APIs, scrapers, and datasets… and met the following challenges:

• Partial coverage and inconsistent time ranges across sources;
• Missing values that reduced statistical validity;
• Significant time lost to troubleshooting and manual data cleaning.

These issues delayed progress and risked weakening the reliability of the results.

One Data365 API, Big Shift

Once the Data365 API was plugged in, Professor Willis finally got what he needed to move on.

What used to be a patchwork of sources turned into a single, streamlined pipeline that just worked:

[[ns-nl]]

1. Keyword-level filtering for tickers and sentiment signals;
2. Access to 91,738 posts from Twitter & Reddit — cleaned, filtered, and complete (just as they were published);
3. 100% historical coverage for the entire 12-month research window;
4. Data that aligned perfectly with Yahoo Finance trading stats

[[/ns-nl]]

This provided a consistent, research-grade dataset, enabling statistically sound modeling and freeing Professor Willis to focus on analysis rather than data troubleshooting.

“Data365 provided exactly what I needed to advance my dissertation research. Before using the platform, I struggled with fragmented and inconsistent data sources, which slowed my progress. With Data365, I was able to seamlessly gather, structure, and analyze a comprehensive dataset, ultimately removing one of the single biggest obstacles in my research journey. The ease of use and reliability of the platform allowed me to focus more on meaningful analysis and less on troubleshooting data collection.”

— Professor Roston T. Willis Jr.

Results that Mattered Or Confidence in Every Number

By integrating Data365 into the research process, Professor Willis:

• Compiled 1,743 daily observations across 7 major companies ($TSLA, $AAPL, $AMZN, $NVDA, $GME, $JPM, $WMT);
• Retrieved over 90,000 sentiment-scored posts with no missing values;
• Created a statistically reliable dataset ready for deep analysis using VADER;
• Developed practical insights that will inform academic papers, classroom exercises, and student-led projects;
• The research is now underway with full confidence in data quality, moving from dataset assembly to meaningful discovery.

By eliminating data gaps, the integration freed the research process to focus on its core objectives: instead of hunting errors, Professor Willis now builds models, tests assumptions, and teaches with confidence.

“I’m pleased to share that I have successfully defended my dissertation, an achievement made possible in large part thanks to Data365. Your platform was instrumental in overcoming the challenges of data collection and ensuring the completion of my research.”

Dr. Roston T. Willis Jr.

DBA, MBA