Crampton’s article on Big Data, its limitations, and possible methods of improving the usability of the geoweb is certainly one of my favourite articles so far. I found the article to be informative, entertaining, and well written.
In a class discussion earlier this semester we discussed whether or not Twitter and other social media platforms were creating usable, interesting data or, if they were creating ‘noise’. This article discusses some of these very issues. How does one find usable data when ‘bots’ are hijacking trending topics to sell iPhones? Is any of the data usable anyway? The fact that Twitter is largely comprised of ‘re-Tweets’ lead the author to conclude their data set was fairly thin.
I think the author made a good point when pointing out that observations cannot be generalized to society-at-large because of the very small usage of Twitter, and the small proportion of Twitter users who actually Tweet. The lack of explanatory power within some of these very large datasets was also identified as a major restriction of using ‘Big Data’, specifically Twitter.
The suggestions of using the geoweb as ancillary data seemed reasonable given the limitations of many Big Data datasets. To me the most interesting representation of the LPS data was the Tweets over time graph. I would be really interested in seeing how other trending topics peak and fall over time.
GoOutside