Principles of a sound golf swing

Here's an article I recommend to anyone looking for the principles to a fundamentally sound golf swing. In particular, just read the forward (page 6) by Hal Sutton's coach, Floyd Horgen.


And a quote from Hal that I wholly agree with: "Most golfers are trying to hit the ball with the clubhead. The
world’s best players swing the whole club and allow the ball to get in the way of the clubhead."
Read more...

Is PowerPoint leading to poor decision-making?

This article on the Pentagon's (over-)use of PowerPoint makes that argument.


While the author acknowledges some good use cases for ppt (info briefings, kick-off lectures, ...), he has experienced ppt's creep into areas where they do not meet the need. Even worse, their use leads to hasty decision-making that has little (to no) traceability.
Read more...

Do you hate meetings?

Paul Graham's latest thesis regarding the "Maker's vs. Manager's" schedules sparked a debate with my colleagues, and apparently with some at the Freakonomics blog.

Your thoughts? Read more...

ASONAM'09 Concluding Remarks

The quality of the conference attendees and talks were a pleasant surprise, particularly considering it was a first-year conference. It was one of the few academic conferences I've attended since entering the professional world where I felt the presenters and discussions were in line with reality (vice overly theoretical). The committee made excellent work of selecting the choice of keynote speakers.

Some interesting work/findings presented:

1. Technology adoption reinforces face-to-face social activity (Putnam & Kolko, Univ. of Washington)
2. Legal predictions regarding which lawyer will go before which judge, a process that is supposed to be random!! (Bartal et al., Israel)
3. The president's job approval is favorably impacted by a split government -- that is, when Congress is held by the opposition party (Danowski, Univ. of Illinois-Chicago)

This is a conference worth tracking over the next few years -- next year's will be hosted by the University of Southern Denmark. Read more...

ASONAM'09 Recommendation Systems Session

Dr. Furtado of Univ. of Fortaleza (Brasil) presented work he is doing in crowd-sourcing to raise the awareness of crime; transparency of crime reports is a particular problem in South America. Using data from wikicrimes.org, he has done work to identify crime hotspots by clustering events by locality and level of participation (e.g., number of events, number of participants, type of crime, ...). This could be extended for a number of useful applications: opinions on places to <add activity here>, infection outbreaks, suicide attacks...

I found many other talks related to recommendation systems, and I extracted some general principles:

1. Identify items/fields of interest in an event
2. Get a feed of events
3. Perform a similarity calculation on events
4. Display recommended results to the user

The interesting work is in #3. Many talks addressed inferring indirect links -- when a user doesn't directly score an event or express a likeness to some other actor. Talks presented applications to concert suggestions, navigational pattern similarity, forum postings and text mining for word co-occurrence. Read more...

ASONAM'09 Day 1 Keynote

Dr. Faloutsos (CMU) did a good job of kicking off the discussion with real-world examples of applicability of graph analysis (food web to determine most important animals in ecosystem, boards of directors to find conflicts of interest, ...). He showed experimental properties extracted from Zipf (log/log) plots of degree vs. rank applied to web links, click-stream data, citeseer and trust data (epinions.com). The slope of these graphs tend to 1.7 and the eigenvalue slope is ~0.8. This, he said, could be very good to identify outliers in connectivity: he applied it to call records to find telemarketers.

His research group has explored how graphs grow over time. They found that the diameter of dynamic graphs shrink (i.e., the graph becomes increasingly more connected). For Autonomous Systems (ASes), the number of nodes increased by 2 in 18 months while the number of edges increased with a slope ~1.7 (more than double); he noted that a slope of 1 indicates a tree (hierarchy) while a slope of 2 is a clique. They explored the non-largest connected component (NLCC) to determine whether is grows, shrinks, or stabilizes over time. And they found...that is oscillates between the 3 states. They observed edge inter-arrival time in LinkedIn data and found that popularity declines in a power-law distribution with slope -1.6 (close to Barabasi -1.5) -- that is, the "rich get richer".

Finally, he has been researching graph generation for 2 reasons:

1. as a data source when privacy concerns arise
2. to understand and enumerate valuable properties of dynamic graphs

They were able to generate pseudo-network graphs that had the properties one expects of the graphs -- and they can be perpetuate as the graph grows using their techniques. Their work can be found in ICMD'08 and KDD'07. He put in a plug for Hadoop MapReduce for graph processing. Read more...

ASONAM'09

The inaugural conference on Advances in Social Networks Analysis and Mining (ASONAM) was the first umbrella conference to focus specifically on SNA from a multi-disciplinary perspective. It was held 20-23 July in Athens, Greece. Read more...