While searching for information related to our Big Data
Course I found a web based data analysis tool called Statwing. This is a free
program that anyone can use to check the statistical significance of one
variable on another variable. For the video tutorial below I choose to look at
NBA player statistics to see if there is any correlation between the team the
player is on or the position they play on the # of games played, minutes
played, usage rate, true shooting percentage, percent of field goals assists,
rate of assists against possessions used, turnover rate, offensive rebound
rate, defensive rebound rate, total rebound rate, and player efficiency rating. I gathered my data from www.hoopdata.com which is an affiliate of
ESPN. In some areas the results kind of surprised me but many others were as
expected. I hope you find the following
video useful. Thanks for watching.
If you would like use this program please go to: www.statwing.com. It is easy to use and just requires an email address and password to sign up.
I read Chris’ blog post and decided to try out Statwing. My first impression of the website and the statistical analysis was that this package seemed very intuitive and easy to use. The interface was easy and it takes only three simple steps to go from a huge spreadsheet of data to direct comparisons of variables to search for significance.
ReplyDeleteI understand the long mathematical process of testing variables for significance, and I have used software packages such as Minitab, which Chris mentions in his video tutorial, and Excel to find significant variables in data. This Statwing’s statistical analysis tool is much more simple to use than either hand calculations or other software packages that I have used, and it calculates the same values. Having a solid mathematical background of calculating p-values and testing for significance made me more comfortable using Statwing’s statistical analysis tool and made me appreciate its power and speed.
For my analysis using Statwing, I chose to use 2012 Major League Baseball stats. We are in the midst of fantasy baseball drafting season, and I was interested in finding out if league, position, or team had any effect on a player’s performance in major hitting statistical categories. This could have an impact on a person’s drafting strategy.
After running the analysis I found some interesting results. The American League is known for better hitting since this league has a designated hitter, so I presumed that being in this league might be an advantage for hitters. Surprisingly, Statwing’s analysis told me the opposite. The National League had a slightly significantly higher batting average, on base percentage, and triples.
Next I checked whether position made any difference in hitting performance. Statwing’s results showed there is a much stronger relationship between position and hitting statistics than league and hitting statistics. Position of a player had a significant influence on the player’s hits, doubles, triples, homeruns, strikeouts, stolen bases, caught stealing, and slugging percentage. The results tell you which positions have significantly higher or lower values. This would help when trying to draft a team.
Lastly I checked to see if a player’s team had any influence on his performance. I thought this would be a valid comparison because each ballpark in baseball is different; therefore a player who plays for a team with a small ballpark may hit more homeruns than a player who plays his home games at a large ballpark. After running the analysis I found that this approach was somewhat misleading. Team was significant for almost all hitting categories. I believe this is because some teams are much better offensively than other teams that there is always a statistically significant difference between them. So a player who is very good can be on a team that is not very good, and his production might not be noticed as significant.
In the end, I found through Statwing’s analysis that position is the most important factor in determining a players offensive production. I thought league would have a greater influence than it actually did. Statwing is a tool that is very easy to use. It does not take much time at all. It is very intuitive and provides valuable analysis. I would definitely recommend this tool.