Social Media for Good Governance: The Quality Challenge

In my last blog, I talked about governments will serve themselves and their citizens better if they do not take an adversarial stand against social media, and instead proactively use social media for good governance. In order to do so, governments have to be able to harvest and use social media data effectively.  For social media to facilitate good governance a fundamental prerequisite is a quality search in the social media space.   

Social media data quality
 
In order to explore the quality issues, let us first look at what we can do with data.  This in turn will help us better define quality of data.

In general, one can do three useful things with data: a) perform searches, b) analyze trends and c) summarize data.  For our purposes, we will look at the a) and b) – the problem of search and trend analysis in the social media domain.  With social media search, we would look for information on a pre-defined subject (e.g. MDGs or Millennium Development Goals).  For social media trend analysis, we may be interested in what the trends are for a particular subject (trend analysis for specified subject), or in an open-ended manner (open-ended trend analysis).

Social media search:  Google search for social media (Update) is an example of a search on a pre-defined subject using specific search terms (see screen shot above).  Here is an example of a Google social media search for “Obama”.

Social media trend analysis:  can be open-ended or defined. 
 
a) For example, with Google Trends, you can track the world’s interest in selected topics (trend analysis for specified subject).  

b) With Google Hot Searches, you can see a snapshot of what’s on the public’s collective mind by viewing the fastest-rising searches for different points of time (open-ended trend analysis).

Now, if we look specifically at social media search, we see that the very nature of social media creates barriers to conducting an effective search. 

The unique properties of social media data that lead to challenges in search

As I mentioned in my last blog, the four characteristics of social media - chaotic, huge, free, and collectively good but individually unreliable create issues with quality and thus make the use of social media data for good governance, difficult.   Here we look at how the four characteristics create challenges:

1. Chaotic:  Social media is by its very nature, fragmented, diverse, and disorganized.  Social media today is widely dispersed and highly differentiated.  It is an amalgam of different services/ websites such as Facebook.com, Twitter.com, YouTube, Flickr, Google Buzz as well as real time feed aggregators such as FriendFeed.  Social media is also available on different platforms such as mobile and PC, both with different set of attributes.  See here for a previous ICT blog on mobile social media and governance.  As an example of the diversity of social media, here’s a wiki table on the kinds of social media sources out there (see Table 1)  
2. Huge:  In the US alone, 75% of the online population uses social media. Social media data is constantly growing – both social networking sites and blogging sites grew by around 50% in the last year and time spent grew by 82% (Nielsen study).  The size of social media, taken as a whole, can be overwhelming.  As of June 2010, there are 65 million tweets per day.  Here is a graph of Twitter usage earlier this year.  Facebook chat usage is currently over 1 billion messages per day. 

Table 1: The wide spectrum of social media (Source: Wikipedia.com)

3. Free:  As there are no filters on social media, it is spontaneous, honest and hard to control (See Table 2 for the number of searches in Pakistan and Bangladesh on the term “Facebook” before and after Facebook was banned in the two countries).  This can make social media data valuable for improved governance but can also pose some problems as there will be lots of data that are incorrect, misleading or spam. 
4. Collectively good but individually unreliable: Partially due to the lack of filtering and controls, there is a large variation in the reliability and quality of the data.  As a collective source, social media data can be reliable, but individual sources vary enormously in quality.  
    

Thus we see that these characteristics create significant challenges in performing an effective search of social media and therefore for using it to improve governance.  The biggest obstacle is the quality, relevance and reliability of the data.  How do you separate the wheat from the chaff?  There are particular technical challenges that we need to address. We will explore some of these issues in my next blog

...................................................

This is the second blog of a three-part blog series on Social Media for Good Governance (Social Media for Good Governance: No Silver Bullet Yet- I, Social Media for Good Governance: The Quality Measurement Challenge - III)

Photo Credit: Flickr user Jenser (Clasix-Design)