Comments for TagWalk Blog

Comment on Experimental Foursquare Recommendations by Tom Rowan

Tom Rowan — Wed, 04 Aug 2010 08:17:02 +0000

Tim,

That’s really interesting stuff. I’ve been thinking a lot about location services recently and I had come to the same conclusion that it would be possible to datamine the location data to produce hot spot graphs or reputation findings.

Are you able to include data about the time that the checkin occurred? (You may be doing this already.)

An application of this might be to find somewhere to eat. If somewhere is popular, then a simple analysis of the data suggests that I might want to go there. I also might want to NOT go there simply because it IS popular at lunchtime on Wednesdays. If we know that a place is popular at a given time, can we infer that it isn’t popular at another time? (I’m not sure whether we can; my gut feeling is that you can’t prove a negative here.) But at least times when a venue is popular might be of particular interest to businesses.

Just a thought.

Keep up the good work!

Tom

Comment on Twitter User Reputation Computed from Tweets by Eugene Mandel

Eugene Mandel — Thu, 14 Jan 2010 19:56:44 +0000

Hi Tim,
Very good post. Thank you for mentioning MustExist in the lists section (and, of course, thanks to Eric for bringing it to your attention). I agree that reputation computation should take into account all signals available: followers, friends, list memberships, RT’s, replies (engagement), and content.
TagWalk looks very promising.

Comment on Twitter User Reputation Computed from Tweets by Tim Hastings

Tim Hastings — Thu, 14 Jan 2010 19:36:56 +0000

Thanks Eric, I have updated the article to mention MustExist’s List Tags – top tip.

I think retweeting is an important part of reputation building because it amplifies the relationships between the target user(s) and the content of the tweet (words, hash tags etc.)

TagWalk for example counts retweets too and also increments retweet counters against the content to build a picture of words, tags, users etc. that are often retweeted.

I think Social Media Reputation Optimization is implicit in any Twitter based marketing campaign, as such campaigns are often designed for retweeting a specific message or incentivising users to post something including a particular tag or phase.

We may also see careful wording chosen for the “tweet this” buttons seen on websites.

Comment on Twitter User Reputation Computed from Tweets by Eric Andersen

Eric Andersen — Thu, 14 Jan 2010 17:25:21 +0000

Great post, Tim. I would love to hear more about your ideas on “Social Media Reputation Optimization” and what type of activity, standards, etc might evolve for this. Also, what about reputation based on retweets as a subcategory of mentions? Apps like Twitalyzer and Klout do this sort of calculation, that seems highly related IMHO.

Finally, just one quick comment – you should mention MustExist in the section on Twitter List-based reputation as that seems to be the main app doing this kind of reputation analysis based on Twitter Lists.

Comment on Twitter Statistics Now Showing on TagWalk by Jon Pear (a.k.a. NeuroAster)

Jon Pear (a.k.a. NeuroAster) — Sat, 04 Jul 2009 18:36:17 +0000

Tagwalk Dot Com is WAY more useful, intuitive, and in4mative than Hashtags Dot Org Your awesomeness rules

Comment on PHP Problem, Failed to Open Stream – File Too Large by Tim Hastings

Tim Hastings — Mon, 22 Jun 2009 14:39:50 +0000

Sadly, the simplest thing you can do is work around this limitation in PHP by working with smaller data files, i.e. splitting the file into smaller files and writing code to work on the separate pieces.
This is by no means a perfect solution, but a pragramtic side-stepping of the problem.

Comment on PHP Problem, Failed to Open Stream – File Too Large by hanke

hanke — Wed, 17 Jun 2009 09:57:15 +0000

so what should i do, coz’ i’m facing that problem right now? failed to read file > 2GB. PLZ help me .. if you have any idea. thx a lot

Comment on Data Mining Twitter, a BarCamp Leeds talk by timhastings

timhastings — Fri, 05 Jun 2009 08:44:17 +0000

Thank you for the sparkline suggestion, I agree, they would be interesting eye candy.

It would require keeping summaries by date which I dont have at the moment so is not something I could add very quickly.
Ill add it to my look into list.

Imran – no, no, thank you! It was a great event and a fun opportunity to make friends and get some feedback on my project.

Comment on Data Mining Twitter, a BarCamp Leeds talk by Imran Ali

Imran Ali — Thu, 04 Jun 2009 10:54:41 +0000

Thanks for coming along fella – I missed your talk, but picked up a lot of positive buzz about it throughout the weekend…thanks again for contributing and attending, hope you enjoyed the other sessions too

Comment on Data Mining Twitter, a BarCamp Leeds talk by andymurd

andymurd — Thu, 04 Jun 2009 09:40:02 +0000

Tagwalk looks like fun, Tim. It’s a shame I couldn’t make BarCamp Leeds this year to see your talk in person but this write-up has sparked my interest.

Can I make a suggestion?

I’d love to see a sparkline against each link showing the interest since it was first tweeted. I think it would be fascinating to see how different types of story grow and fade, or am I being foolish and all the sparklines would look the same?