Social networks are fascinating by their rapid growth and their aiblity to provide openly accessbile data.
What it does
The Beaker notebook examines the relationship between stock markets and the average emotional score (positive is 1, negative is -1).
How we built it
We collected and compared the data of 8 companies like Amazon, Google, Netflix both from the NASDAQ API and the tweets that mentionned these names. Some statistical tests were conducted to assess the correlation levels.
Challenges we ran into
The Twitter API has some serious limitations for this kind of task (e.g., timed queries)
Accomplishments that we're proud of
Although not new, the idea to use social networks as approximation to real industries can help researchers quantify the confidence with which they rely on alternative sources of information.
What we learned
In many cases, phone applications can be useful for some specific tasks. However, as this project has shown us, web apps and their APIs tend to be faster, lighter and easier to access for developpers and/or data scientists.
What's next for Feeling Big Data:sentimental analysis of stock markets
Although only a proof of concept for now, the stock market sentimental analysis can be conducted even further by fitting a Hidden Markov Model, often used in time series analysis and forecast, to assess its performance on the twitter data.