Search Engine History
Infographic by the PPC Blog.com
Tag Archives: Visualizations
#slaname Social Network Graph
Social network graph of #slaname tweet replies October 14, 2009 to December 11, 2009.
The more lines you have, the more @replies to different people you sent. If you don’t appear on the graph, but know that you sent out @replies, it’s because the person you sent your @reply to never sent out an @reply and so that person won’t appear on the graph and unfortunately, you can’t either!
For a more detailed explanation of how this was done, see #sla2009 Social Network Graph.
Based on the code of eskimoblood.
Created using Processing with data from the #slaname Twapper Keeper archive setup by iBraryGuy.
For larger sizes, see the original on flickr.
#sla2009 Social Network Graph
Using Processing and sample code from www.eskimoblood.de I have created a basic social network graph of replies sent during #sla2009.
The thicker the line the more times you sent an @reply to that
person. The more lines you have, the more @replies to different people
you sent. If you don’t appear on the graph, but know that you sent out
@replies, it’s because the person you sent your @reply to never sent
out an @reply and so that person won’t appear on the graph and
unfortunately, you can’t either! Interestingly, a few people only sent
replies to themselves, so they do appear on the graph as a line that
goes back to themselves.
The sample code from eskimoblood required that my data be in the following format:
- One array of all the usernames in alphabetical order
- A second multi-dimensional array that links the users with each other (in this case based on @replies) using an ID number. Like this:
@amalthea67
{70, 74}
In other words @amalthea67 sent @replies to the 69th and 73rd person in the list of usernames on the wheel. The first person in the list is at number 0.
Here is an outline of what I did:
- Queried the #sla2009 re-tweeter database for all tweets with @replies as of June 19, 2009
- Using php, created an html table of this query that lists the ID of the tweet, the username of the person who sent the reply and the username(s) of the person(s) it was sent to. Here is a screen shot of part of the HTML table that resulted.
- Imported the html table into an Inmagic database, which allows for multiple entries in the “Replies” field (separated by a space)
- Created a secondary Inmagic database for every username, plus a unique ID for each.
- Exported the data to a report that lists the username and the ID number of the person the reply was sent to. Here is a screen shot:
- Using Excel again, this table was copied, pasted and transposed so the usernames ran along the the first row and the numeric ID of the target of the @reply ran along in the second row. A little bit of find/replace to get the data in the exact format (as above) and voila! My data is in the format that I need it in.
- I made a few changes to eskimoblood’s code and ran the sketch in Processing, which resulted in the social network graph above.
Update: this graph is now available with color.
BloggingStream by 