Twitter for drug trend monitoring

Today I participated in the workshop Making sense of Twitter at the Communities and Technologies conference in Brisbane. My contribution was simply the idea that Twitter could be used in drug trend monitoring, as explained below:

Some people are using the internet to seek drug-related information, share their drug use stories with like-minded others, and buy pharmaceutical and novel substances marketed as herbs or ‘legal highs’. Researchers have responded to this trend by conducting monitoring studies, including tracking websites that sell psychoactive substances (e.g., ‘Psychonaut project’, Schifano et al., 2006) and analysing the contents of online discussion among people who use drugs (e.g., ‘Real drugs in a virtual world’ project, Murguía, Tackett-Gibson, & Lessem, 2007). Apart from Lange et al.’s (2010) study of YouTube videos depicting young people using the hallucinogen Salvia, there has been little research into the use of social media to share drug information and to advertise websites that sell psychoactive substances.

Using 140kit, I collected Tweets that contained the word ‘mephedrone’, the name of an amphetamine-type substance that was banned in the UK in 2010 and is banned under the Analogues acts in Australia. 360,755 tweets were posted by 217,739 Twitter accounts in one week during March 2011. Browsing through a small selection of these Tweets indicated that most of them purport to sell mephedrone and include URLs to vendor websites, while smaller proportions could be categorised as news, policy and research about the drug, as well as casual discussion among people who appear to use the drug. I can see the potential for a number of studies using these data from Twitter. Drug trend monitoring could be enhanced by setting up a system to store and count the number of tweets containing drug terms, tracking trends over time. Tweets could be scanned for ‘new’ drug terms for which new searches and monitoring may be instigated. Geolocation may assist in understanding the potential for that drug being used in specific parts of the world. Hashtag analysis may indicate social networks of people who discuss these drugs. Analysis of Twitter could feed into current attempts to monitor drug vendor websites, as the purpose of most Tweets is to encourage sales. Through this collection, I also stumbled across a video-sharing site designed to share videos of ‘funny tripz’ using ‘legal highs’. Monitoring sites like FunnyTripz and YouTube could also assist in drug trend monitoring, especially in tracking the existence and use of new legal highs that are yet to come to the attention of health and legislative departments.

So, what did we learn in the workshop?

Firstly – I learnt about this tool called Pirate Pad, which is like a public whiteboard that anyone can edit – what a fantastic tool for collaboration. For our Pirate Pad for this workshop, see http://piratepad.net/B9kg8sNSGL

I will save a version of this piratepad in case the link fails later down the track, but basically all the links are there. Axel Bruns and Jean Burgess at Mapping Online Publics shows us their process of extracting twitter datasets, cleaning and preparing them, and then analysing and visualising them.

We also heard from a selection of researchers with alternate approaches to Twitter analysis, including Cornelius Puschmann‘s research group who introduced the ideas of small vs big data, and emphasised the importance of qualitative analysis to inform quantitative and network analysis of (what are often) huge datasets.

Aneesha Bakharia from QUT also spoke about algorithms for the thematic analysis of twitter data. It appeared to me that her tools identify underlying themes using a combination of quantitative analysis with qualitative inspection of the results and therefore could be used with large scale data.

I now feel armed with a set of freely available tools (licensed Creative Commons, thanks Axel!) to play around with the idea of using Twitter for drug trend monitoring and ultimately to fund a project to make this happen. And there is also a network of potential collaborators to help me do it.

Yay! 🙂

 

Websites
140kit mephedrone tweets – http://140kit.com/tronica/collections/3028
Funny tripz – http://www.funnytripz.com

References