Embedded Files
pyTSA - A Python Time-Series Analyzer (pronounced as "pizza")
Analysis of dynamical systems is often done via simulation ensembles, under different parameter configurations, or by repeatedly sampling from a random process. pyTSA is a Python tool to make analyse of such time-series automatic, its scripts can be pipelined with any simulation tool outputting time-series, and intuitive commands allow to perform complex analyses in a intuitive way.
Input and Output data
Input and Output data
A dataset is a set of files in the usual csv/tsv/.. format; each file is interpreted as the result of an experiment, e.g. a model simulation, where each column is a variable of your system. pyTSA has commands to load datasets and assign mnemonic names to model variables (columns). Plots can be exported in png/eps/raw/.. formats.
A dataset is a set of files in the usual csv/tsv/.. format; each file is interpreted as the result of an experiment, e.g. a model simulation, where each column is a variable of your system. pyTSA has commands to load datasets and assign mnemonic names to model variables (columns). Plots can be exported in png/eps/raw/.. formats.
Can I see my data "as it is"?
The pyTSA data analysis pipeline
Yes. You can plot as many time-series as you want from a dataset, possibly displaying each system variable in a separate panel. You can also plot such variable in 2D and 3D phase spaces
What is the "average" behavior in my data?
You can ensemble different experiments and estimate a "average" trend by averaging a set of time-series, and looking at the standard deviation of each variable. This is the simplest of all the aggregate measure implemented in pyTSA.
What is the probability of finding X units of chemical Y after Z time units?
If your model is stochastic you usually want to know what is the empirical probability of its state variables at a certain time instant.
Once you have sampled your model state as many times as you want, you can estimate such density function and fit it to a Gaussian with pyTSA,
What is the probability of finding X units of chemical Y over time?
One might be interested in estimating the probability of observing certain units X of chemical Y all along the model execution.
In stochastic models, this is equivalent to assess the empirical solution of a master equation. pyTSA performs that via 2D heatmaps or 3D surface plots.
The current version of PyTSA is hosted on Github.
Automatising the analysis of stochastic biochemical time-series. G.Caravagna, L.De Sano and M.Antoniotti.
BMC Bioinformatics 16(Suppl 9):S8, 2015.
Google Sites
Report abuse