Concept

In this experiment, approximately four million Tweets originating from Ireland were collected between September 21st and December 14th 2010. Tweets are short, information-rich messages that people around the world (including Ireland) post to Twitter. I mined Twitter and built algorithms that tagged the Tweets as being positive, negative or neutral. In doing this, I created a very accurate picture of how the emotions and moods of people in Ireland change and fluctuate over time.


Video

This is a time-lapse video of how the collective moods of Ireland fluctuate over an average day.



Weekly Analysis

A graph detailing the collective mood of people in Ireland changing over an average week:

From this graph, we can see that there is a definite dip in the mood during the middle of the week. People are happiest on Friday at about 18:00, and the trough (most negative mood) occurs during the early hours of Thursday morning.


Daily Analysis

A graph detailing the collective mood of people in Ireland changing over an average day:

From this graph, we can see that on average, most people in Ireland are happiest at about 18:00, whereas they're least happy at about 04:00. The blue line represents the overall rise in the mood over the average day.


East Coast vs. West Coast

This graph compares the general mood of the East Coast to the general mood of the West Coast of Ireland over an average day. From this graph, we can see that the West Coast of Ireland is almost always happier than the East Coast.


Budget Day - December 2010

This graph compares the mood of an average day to that of the Budget. At about 17:30, the collective mood of the people in Ireland begins to plummet rapidly, this correlates with the time the Budget was being read. This sudden drop in the mood is consistent for a few hours, therefore indicating that it is not an anomaly. Indicating of course that people weren't too happy with the Budget.


Ireland agrees to request financial support from the EU - November 2010.

Another very interesting graph. During November 2010, the Irish government agreed to request financial aid from the EU. We can see from the above graph that for the most part, the mood that day was well below average. However, around the time the Government announces its plans, the mood soars upwards, indicating that generally, people were quite happy / relieved about the news.


Report

I wrote a report on this project to detail how I did it in much more detail. You can view it by clicking here.


Real Time

Real Time mood tracking is also a big part of this experiment. I created a series of algorithms which continually track the sentiment of Ireland and reflect the mood on a auto-updating map of Ireland, whilst graphing the change in mood every hour.


What I used

In doing this experiment, I used the following. I wrote the algorithms in Python, a programming language well suited for this task. I also built a data miner using PHP, to mine tweets from Twitter. To rate each tweet, I used the Subjectivity Lexicon, courtesy of The University of Pittsburgh. To sort the tweets into the different counties the came from, I mainly used my own algorithms but also used the Bing Maps Geocoding API for a certain portion of the tweets. To make all of the graphs, I used Omni Graph Sketcher.


Coverage


Awards


About / Contact

I am a nine-teen year old student. You can contact me at james@jameseggers.com for any queries you may have. This Project was exhibited at the BT Young Scientist and Technology Exhibition 2011.