Brand Tags Data Mining
Hey everyone, sorry for the interruption (and slow posting lately), just need a favor. I have to do a presentation next week about Brand Tags and I’m trying to find some interesting new data to pull and share. What sort of stuff would you want to see? How would you like to use the site? Also, any additional feature ideas are welcome. Thanks a bunch.

Hi, I'm 
I would be interested in correlations between brands. Does a brand have 20 common tags with another brand? Any interesting insights there?
I’d like to see all the brands associated with a particular word. Also a table that sorts brands with negative sentiments against those with positive sentiments. Converting Brand Tags from a process associated just with sentiment into something that mines that sentiment and converts it into data – that’s what will keep people coming back.
How about tags for a particular brand over time? Would be interesting to see the ebb & flow of how people were tagging brands. Is there seasonality for brands? “Better” tags for Sony pre-Christmas rather than post-Christmas? What about events, like the Superbowl? I’m thinking the UI would be similar to what Delicious does for search now or even what Wikirank does. Maybe these URLs will work: http://delicious.com/search?p=hadoop&chk=&sd=&ed=&fr=del_icio_us&lc=0&atags=hadoop&rtags=&context=all%7C%7C
http://wikirank.com/en/Hadoop
Looking forward to it.
I would want to see how long people hesitated before submitting a tag. Could speak to the clarity of each company’s brand message…or, in the case of a controversial brand, it could help describe the severeness of opinion polarity.
Along the same lines, it would be good to know how long people waited before clicking “skip”.
i.e.: how much time does a brand have to convey its message before someone decides it’s not clear enough to articulate.
something basic — the number of different associations per brand — lots of different associations might be an indicator of richness of brand perceptions, but might also point to lack of clear brand identity
p.s. i like dave’s suggestion also — common words
i wonder if you could create a category map to get a quick snapshot of a particular industry by selecting a competitive set of brands? Without getting too complicated, perhaps there’s a way to view on an axis or perceptual map.
I would try to link the results of brand tags with tweets about the brand.
Why not thinking in the sense of a network of atributtes surrounding a brand, acognitive map of the brand?
You’ve heard a few of my suggestions before, but I’d echo the calls for a presentation of data that incorporates variations on the timing theme:
1) Is the severity (positive and/or negative) of tags directly or inversely related to the time it takes someone to tag?
2) A time series analysis of a brand’s tags would be great. Even better would be a mapping of major events (news stories, campaigns, etc) onto the time series to see how events affect perception. How immediate is the shift (if there is one) in sentiment? What does the lag look like? How long does it take to regain positive perception (if you can)?
Also, I think understanding who is doing the tagging is important.
3) Correlations between referral pages and tags, possibly the type of sentiment and the referral page. Were swift, negative tags associated with certain referral pages for particular brands or industries? Etc.
I know the presentation is over and all but for whenever you do your next presentation I think an interesting future study may be Microsoft/Windows 7 before and after the launch of the new operating system.
I’d like to see a more international focus. Some brands have a completely different perception in the US than they do in Europe and I think it’s interesting to see how perceptions vary from country to country.
This is funny. I was just yesterday afternoon talking with a colleague and saying how much I would love to see Brand Tags for a time series, especially for industries such as financial services and automotive. Maybe you could choose a few brands and the top 20 tags could show over time with a number of responses. Somewhat similar to Google Trends which gives a comparative graph. Thanks Noah, great work as always.
Would be interesting to see if there is any tendency in emotional vs. rational associations.
Adding gavin’s suggestion of tags over a course of time, wonder what brand tags will be like with Google motion charts (showing the top tags), and associated with google news (just like oogle insights does to the search chart).
Just want to drop by and saw I enjoy the presentation very much. As a branding manager for my rugs company, it makes ponder about the future of brands using tags.