In this section we have made available a number of white papers that give an insight into the type of analysis that we can undertake. Feel free to download and view any of the papers that are of interest to you. If you would like a particular topic covered, let us know and we will consider doing a piece of analysis on it.
We have looked at analysing tweets made over the weekend of the 6-8th August following the riots that occurred in London and then spread across England.
We downloaded tweets and then analysed them using our contextual analysis techniques to see whether twitter was in fact used to direct and organise the gangs activities, in terms of where they were going to attack next. The results appear to indicate that Twitter was in fact used to co-ordinate attacks, perhaps in parallel to the Blackberry network. The whitepaper can be seen here.
We have produced a whitepaper that shows how Brand Aura's discover product can be used to analyse and further investigate the recent T in the Park music festival.
We downloaded all tweets made over the weekend of the festival, and could use them to determine who had been the most popular act, but more importantly we could give an insight into why the public liked each particular act. The whitepaper showing this analysis can be seen here.
We have produced a whitepaper that gives an overview of the problems facing the creative industries when trying to understand a new market. For example, this is required when developing a new advertising campaign or for brand development.
The whitepaper describing how our analysis is perfectly suited to this task can be seen here.
Over a period of three months from January 2011 until end of March 2011 we downloaded any and all tweets related to the financial sector. We conducted contextual analysis on a daily basis as the volume of tweets was quite large (of the order of over 10k tweets per day).
The volume of tweets allows us to examine how the public perception of key themes in the financial sector changed over time. It also allowed us to undertake some competitor analysis beyond simple sentiment analysis. The results and a discussion of them can be seen here.
In April 2010, BP suffered a huge oil spill known as the Deepwater Horizon oil spill that resulted in millions of barrels of oil being leaked into the Gulf of Mexico. More information on the spill itself can be found at the Wikipedia article here.
We undertook some analysis of the tweets that were made during this period, and a whitepaper that highlights the contextual analysis and the changing trend of words that were coming up against BP can be seen here.
In May 2010, the UK held its general election to elect members to the House of Commons and form a new government. Information about the result and the campaigns undertaken can be seen at this Wikipedia article.
We analysed a range of online data sources, including the BBC's Have Your Say, forums on the Daily Telegraph website, and football fans forums. In each case we looked to give an insight into the particular hot topic of the day, and we were able to predict a number of decisions that were made by the general public or the political parties, and more importantly give the reasons behind these decisions.
The whitepaper can be viewed here.
In the winter of 2010, a number of reality TV shows in the UK were coming to a conclusion. In particular, the UK edition of X Factor, which ran for 10 weeks and which had weekly evictions of contestants based on a public vote. More information and the weekly results for the competition can be seen at the Wikipedia article.
We analysed Twitter as a single data source, with at times up to 1 million tweets being analysed over a single weekend. An example of the type of analysis can be seen in our whitepaper, with further analysis and weekly predictions on our blog.
Our analysis was successful in predicting the winner and also Rebecca Ferguson as finishing second (which went against popular wisdom and the gambling odds at the time). However of more importance is the fact that we could explain why the public voted the way that they did. For example, we could explain from our analysis why Wagner stayed in as long as he did, why the contestants who were voted out were in fact voted out, and why Rebecca did not win but finished second.
We also undertook analysis on the UK version of Strictly Come Dancing, and again successfully predicted the winner although again more importantly we could explain the reasons why. We also used our process to analyse the BBC Sports Personality of the Year and again successfully predicted the winner of the competition. All of this analysis can be seen at our blog.
We thought a good example of how our process works would be to work out what is the "best" whisky. This is not straightforward - and quickly we realised that like many things this is an incredibly subjective question. What do we mean by best? What criteria should we use to work out which whisky is better than any other?
In the end we decided that we would look at the language that the whisky companies themselves used to describe their product. So words like: smooth, golden, malty, peaty. So the question then becomes what whisky is closest in context to these words. Other words could be used of course, and this would give a different answer. But using the words that the companies themselves use gives an indication of how successful the companies are at using the right language in communicating to their customers.
Our analysis was undertaken over two months, so that changes could be tracked over time. You can see the whitepaper here.
We were asked to undertake a piece of analysis on the perception of the Highlands in the context of Inverness, known as the "Gateway to the Highlands". For this analysis we used tourist comments from the Trip Advisor website.
This analysis highlights the importance of the words that are used when assessing the sentiment of comments, and how the same word can have a positive or negative sentiment depending on its context. The white paper discusses this topic and how our analysis can help in identifying which words should be used when assessing sentiment.
