So let me take you through the steps of creating a DigiVidual.We start with writing up a set of keywords that as a group can be seen to represent the target group. The algorithm has intricate levels of using these words, for example, it gives priority to what we call ‘core characteristics’ – general words that represent the target group at their core, e.g. ‘trust’, ‘connected’, ‘individual’. To build on these, we further include sets of more occasional words linked to these. One of the key ways in which we ‘humanise’ our bot is to also programme routines, hobbies and locations – a digital map of for example, ‘orange juice’ at 9am, ‘football’ at 5pm on a Saturday, and work locations for 9-5 on week days.The keyword combinations then go through various social media sites to discover microblogging, image and blog posts, shopping items and uploaded videos with these words or word clusters attached to them – for this project we included China-specific sites such as Sina, Tudou, Yupoo and Douban. Further, a wider extended search uses both the original keywords and new words the original search has ‘learned’ – words typically associated with the words and word clusters we’ve set them. In this way, the algorithm goes beyond a word search and into deeper associations that we would not necessarily have known of beforehand.The social media content found is then posted on a blog – effectively, an online diary of our DigiVidual, representing the kind of items our target group would be blogging about based on their attributes.Finally we look at the blog and using classic qualitative analysis unpick themes that tell us about the life, aspirations and frustrations of the DigiVidual. And the next step is where it gets interestin
In order to examine the split between "West-leaning" (pro-Western culture) and "East-leaning" (more traditional) youth, we created one DigiVidual for each group in 9 cities throughout China, creating a mixture of tier 1, 2 & 3 cities.The DigiViduals ran for 4 weeks from the beginning of January 2012 to February 2012 and discovered more than 32,000 pieces of web content from multiple online data sources.Not to worry – I won’t attempt to go through all 18 DigiViduals with you today… However, because of the narrative output I mentioned earlier, I should be able to introduce you to one them in just two to three minutes... Mu. I will now show you the Digivideo of Mu – or, the narrative that we saw emerge from the web content blogged to us by the Mu DigiVidual.
Robots journey to_the_east v5
the largest database of human experiences, ever
CHINA… A BIG PLACE too big for qualitative research?
HOW DigiViduals WORK ™1. 2. ALGORITHM COMBINATIONS OF PROGRAMMED KEYWORDS WITH SEARCHED FOR KEYWORDS IN SOCIAL MEDIA4. 3. SOCIAL MEDIA EXTENDED SEARCH USING CONTENT ORIGINAL KEYWORDS POSTED ON AND ASSOCIATED BLOG PHRASES 5. ANALYSIS