Experimental TimeSeries look and feel

timeseries.jl

julia> spx
15914-element TimeSeries Array:
 1950-01-03  |  16.66 16.66 16.66 16.66        
 1950-01-04  |  16.85 16.85 16.85 16.85        
 1950-01-05  |  16.93 16.93 16.93 16.93        
 1950-01-06  |  16.98 16.98 16.98 16.98        
 1950-01-09  |  17.08 17.08 17.08 17.08        
 1950-01-10  |  17.03 17.03 17.03 17.03        
 1950-01-11  |  17.09 17.09 17.09 17.09        
 1950-01-12  |  16.76 16.76 16.76 16.76        
 1950-01-13  |  16.67 16.67 16.67 16.67        
 1950-01-16  |  16.72 16.72 16.72 16.72        
 1950-01-17  |  16.86 16.86 16.86 16.86        
 ⋮                                             
 2013-03-19  |  1552.10 1557.25 1538.57 1548.34
 2013-03-20  |  1548.34 1561.56 1548.34 1558.71
 2013-03-21  |  1558.71 1558.71 1543.55 1545.80
 2013-03-22  |  1545.90 1557.74 1545.90 1556.89
 2013-03-25  |  1556.89 1564.91 1546.22 1551.69
 2013-03-26  |  1551.69 1563.95 1551.69 1563.77
 2013-03-27  |  1563.75 1564.07 1551.90 1562.85
 2013-03-28  |  1562.86 1570.28 1561.08 1569.19
 2013-04-01  |  1569.18 1570.57 1558.47 1562.17
 2013-04-02  |  1562.17 1573.66 1562.17 1570.25
 2013-04-03  |  1570.25 1571.47 1549.80 1553.69

# how quickly can the kurtosis of closing prices be calculated?

R> kurtosis(Cl(GSPC))
[1] -0.5264019
R> system.time(kurtosis(Cl(GSPC)))
   user  system elapsed 
  0.084   0.002   0.085 

julia> kurtosis(Cl(spx))
-0.526626137822463

julia> timetrial(kurtosis, Cl(spx), 100) / 0.085 # 0.085 is the time it takes R to do this
0.04155142270588236

# both Julia and R get a similar result, and Julia is 24x faster