2015 Workshop 6: Qualitative Coding of Large Scale Textual Data

Abstract: Not all twitter data can be meaningfully mined through automation; even sentiment analysis is somewhat imperfect. At times qualitative coding ot Twitter data makes more sense. However, there are a number of ways to set up this workflow. Some are cheaper, some are faster and some are more reliable. In this workshop we will cover coding on Crowdflower using both crowdsourced labor [CSL] and skilled coders. This workshop will cover two different coding efforts on Oxford’s ‘Real Names’ project. The first was using CSL and the second using MSc students trained in qualitative work. We will discuss differences in cost, speed, logistics and strategies. We discuss potential sampling strategies to keep in cost, and time permitting, demonstrate how to pipe data from Netlytic directly into Crowdflower for coding.

Pre-Workshop Prep: 

  • Laptop and power cord

Bernie HoganPresenter’s Bio: Dr Bernie Hogan is a Research Fellow at the Oxford Internet Institute, a department at the University of Oxford. His work sits in between social theory and methodological advances using big data. Concerning theory, Bernie’s work focuses on the ‘exhibitional approach’ and third party agents. That is, how do agents such as Facebook and Twitter mediate our relationships to others? To what extent to algorithms, design and politics intervene on behalf of these agents? Methodologically, Bernie has worked extensively on capturing, collecting and visualizing social graphs, such as Facebook and Twitter social networks as well as sociocognitive networks using cutting edge computer assisted interview approaches. He is published widely across a number of disciplines, although he still considers himself a sociologist. His work can be found at his Google Scholar page: https://scholar.google.co.uk/citations?user=NVU663QAAAAJ&hl=en