Slashdot Mirror
New Social-Network Mapping Tools Compared
Roland Piquepaille writes "There are many new visualization tools around us which try to map our social networks. In this column, I examined Inflow, a datamining tool digging through your email repository to discover and find trends to know more about your networks. Here is a quote: "Assuming you have a significant amount of e-mail traffic, the software will create a remarkably sophisticated assessment of your various social groups, showing you not only their relative size but also the interactions between different groups." I also peeked at TouchGraph GoogleBrowser, which uses Amazon or Google Application Programming Interfaces (APIs) to visually describe how books and Web sites connect with one another. Finally, I took a look at a brand new way of visualizing Google search results, from anacubis. If you know about other similar new tools, please tell me and I'll gather your comments in a future story."
In a recent (i believe 2 months ago) Dr. Dobbs there was an article about just this type of application. There was an article written by one of the top social enginners of applications like this. He was one of the people responsible for doing the Amazon "like this you'd like that" feature.
Anyway, this seems to be a next step in the evolution of search engines, not giving URLs that matches queries, but relating them, showing the relationship between actual data and ubication in internet.
It isn't quite the same but with a little scripting on your end you can make some pretty detailed stats of all your mail using "mls".
--- I do not moderate.
http://marki.host.sk/MLS/
- El riesgo siempre vive - Private J. Vasquez
This was a truly scary demonstration of this kind of technology being used by private industry, namely casinos, to track relationships between people.
Real stream available at: rtsp://media-1.datamerica.com/blackhat/bh-usa-02/v ideo/BH-USA-02-JEFF-JONAS.rm
Revolutions are never about freedom or justice. They're about who's going to be top dog. -- Kilgore Trout
Thank you Roland for the write-up on InFlow on your weblog!
Unfortunately there are some errors...
1) I am not a former IBM'er... they were my first major client.
2) It did not take me 15 years to write the software... the first working version [w/o visuals] was written in 2 weekends in 1987... on a 512K Macintosh... using Prolog. Yes, now it is commercial, used mostly by management consultants, on Windows. I also use it with VPC6 on my Powerbook.
3) InFlow can process data from email traffic to find patterns and paths, but the paragraph you quote is about the OTHER product in the article -- MIT Media Lab's "Social Network Fragments" -- a very cool tool.
Looking at just your own email[in/out] will not tell you much [except that it is 40% spam]. You need to look at the email flows between project team members, co-workers, communities of interest, etc. At least 20 participants before interesting patterns emerge...
Most of our data is collected via on-line surveys -- people participate knowingly. Most survey participants are very eager to see the resulting maps -- they want to see where they, and their friends ended up.
Valdis
http://www.orgnet.com/
http://www.anacubis.com/index.html
http://www.touchgraph.com/TGGoogleBrowser.html
http://www.smartmobs.com/index.html
http://smg.media.mit.edu/projects/SocialNetworkFra gments/
http://www.firstmonday.dk/issues/issue7_4/krebs/
http://www.technologyreview.com/articles/wo_camero n030703.asp
Unaccountable leaders are masters, and unrepresented people are slaves. How do US and EU fare?
With an open source tool instead.
Etherape
Can't believe the author left that one out.
It already is being used. Network analysis (NA) and visualization tools are being used in parts of the government to look at everything from criminal organizations to terrorists. The three biggest hurdles for the government (or for that matter a company that wants to use NA tools for commercial purposes): -- Computing resources and visualization techniques. Only fairly recently has computing power, storage, and graphics capability have gotten to the point where you can deploy these tools in the hands of the analyst. Having the analyst go to a specialized computer terminal/operator is a barrier for its use. The big problem that remains is how to effectively display complex networks. When I looked at moderately complex networks, it quickly became a plate of spaghetti. -- Data sources. Unless you have a common data source, NA tools are exceedingly difficult to implement. You need to capture the entities that are involved, the relationships between, and a indicator of data quality. That data source needs to be integrated with the other tools the analyst uses. For example, if a marketing analyst is reading a survey on who are trendsetters for teen fashions, that analyst needs to be able to seamlessly update a database on who the trendsetters are. That same database needs to feed a NA tool that can be used to determine who the key trendsetters are. -- Analyst rut. Some analysts either have a resistance to adopting new tools or are ignorant of the tools. I think often the problem lies on the tool building side of the house. It is not unusual to see a new application get deployed with little or no input from the users. In one of my jobs, I saw an application get deployed that was completely unusable by over 90% of the users. Apparently the developers were using (new at that time) Pentium-based computers when all the regular users had 486s. The application brought the 486 to its knees. It really is pretty neat when you start with a sparse network with uncertain links and after working it on it for awhile you end up with a solid network, identified new relationships, and discovered new entities.