JLAB_HIGHLIGHTS Introduction to some of the most useful routines. JLAB_HIGHLIGHTS Highlights from the JLAB toolbox. TWODSTATS, TWODMED, TWODHIST. Examining temperature mean and standard deviation over a lat/lon grid, or current speed as a function of say temperature and salinity, or any analysis of multivariate datasets, is faster and easier than you ever imagined. If you are working with data and not using these routines, you should seriously take five minutes right now and try them out. As of Matlab 2015b, these take advantage of improvements in Matlab and are even faster than before, and now support parallelization as well. MSPEC. If you analyze time series, you need a method you can rely on. There are many compelling reasons to prefer the "multitaper" method. In this method, a set of estimates from optimally concentrated tapers or window functions are averaged together, reducing variance while minimizing bias. MSPEC makes using the multitaper straightforward. Just call SLEPTAP first to generate the tapers. MSPEC features support for cross spectra and rotary spectra, as well as 'adaptive' estimation. WAVETRANS. A lot of work and original research has gone into this continuous wavelet transform routine, which I believe is the best you will find anywhere, and it's free. Features natural integration with the generalized Morse wavelets (see below), choice of conditions at the signal endpoints, support for multiple wavelets or mulit-component data, and convenient handling of real-valued or complex-valued data. MORSEWAVE. A detailed analysis of types of continuous wavelets (Lilly and Olhede, 2009) points to a super-family, the generalized Morse wavelets, that should answer your questions about which wavelet to employ. Pick the generalized Morse wavelet with parameter gamma=3 unless you have a very good reason to do otherwise. This wavelet is 'like' the Morlet in spirit, but without the Morlet's serious flaws at narrow time-domain settings---which is when you most want to employ a wavelet anyway. Use MORSESPACE to easily determine frequency bins. RIDGEWALK. If your time series data contains signals that oscillate but also change in time, the wavelet transform can be used as the basis for a remarkably powerful analysis that connects time-varying properties to spectral structure. The starting place for this analysis is RIDGEWALK. This advanced algorithm incorporates original research from Lilly & Olhede (2010a+b, 2011). A key innovation is the ability to view *multivariate* and univariate time series within the same framework as modulated oscillations. POLYSMOOTH. Need to map scattered data, for example, temperature as a function of X and Y? POLYSMOOTH, short for "Polynomial Smoothing", is a powerful and quite general mapping algorithm. Features choice of order of the fit, choice of local weighting function, support for spherical geometry, options for constant fit radius or constant number of data points with variable radius. This algorithm was implemented for the forthcoming Aquarius Sea Surface Salinity satellite. JGRAPH. Finessing figures in Matlab can be a total hassle. Which would you prefer to type, version (a) or version (b) ? (a) z=peaks; plot(z(:,1),'linewidth',3,'linestyle','-','color','b'),hold on plot(z(:,2),'linewidth',2,'linestyle','--','color','r') (b) z=peaks;plot(z(:,1:2)), linestyle 3b- 2r-- They are identical. JLAB's LINESTYLE will save you a lot of time. It includes natural syntax for grayscale, as in "linestyle 3k- 2G-.", where the letter G stands for the shade of gray, between A and K. The JGRAPH module contains many similar time-saving functions, such as YLOG and YLIN, YOFFSET, FLIPY, OUTTICKS, FLIPMAP, etc. PACKFIG. Are your multiple subplots too far away from each other? Don't like having to set the axes properties by hand each time? Try this function. For example, "packfig(3,4)" will squish your 3 rows and 4 columns so they are right next to each other, removing the redundant x- and y-axes labels. __________________________________________________________________ This is part of JLAB --- type 'help jlab' for more information (C) 2011--2015 J.M. Lilly --- type 'help jlab_license' for details