Data Analysis Software
Bios Analyzer.
WHAT IS BIOS ?
You can find out by reading 'bios and the creation of complexity' at chaos tutorials website. Short version: a process that is causal and creative (not creative as in Leonardo Da Vinci, but creative as in time series of a process show features like novelty). Creativity was found in many time series including Schrodinger wave equation, heartbeat intervals, prey predator models, distribution of galaxies. Bipolar feedback (co-existence of opposites) seems to be necessary for creativity. Bios Analyzer is a software that implements methods developed to analyze and measure creativity in time series data.
IS BIOS USEFUL FOR ME ?
The only way to find out is by learning more about it, and by experimenting with this software, as many others did.
Free nonlinear time series data analysis software written in Python.
Currently free as in free beer, soon will also be free as in free speech (as soon as I find some time to refactor the code, and put some comments in it).
If you use this software for academic research, please cite 'bios data analyzer' paper, on which this software is based.
BIOS ANALYZER MAIN FEATURES:
- Analyzes hundreds of data files with one button click
- Handles long time series with more than a million data points
- Plots results and saves them to a tab delimited text file
- Statistics generated when many files are analyzed
- Multiple file menu selects files from different directories
- Simple and easy to use Graphic User Interface
- If you often analyze many files, this application can save you months
- GUI developed with wxPython, algorithms optimized with Numpy, executable made with Py2exe, and all these use Python
The latest version is released in December 2006.
Exact (isometry) recurrence plot of air temperature. | A detail from the recurrence plot on left. | Local diversification's mean and standard error for 8 air temperature series (average length 26575 data points) Analysis done in about 90 seconds and 1 button click. |
For more screenshots, scroll to the bottom.
TESTS:
- Min, Max, Mean, Standard Deviation (S.D.), Contiguity
- Recurrence and consecutive recurrence as a function of radius - test for causality (150 KB doc file)
- *Recurrence plot of series, and multi-dimensional series embedded with the Euclidian norm (isometry recurrence)
- *Embedding plots - recurrence quantification analysis (RQA) on a range of embedding dimensions: isometry; consecutive isometry; recurrence/novelty; arrangement (non-random complexity)
- *Local diversification - increase in variance with the increase of the length of the sub-series
- *Process entropy - a measure of diversity and symmetry of the series
- 95% Confidence Interval Standard Error (S.E.) for the tests above
- Any other test that you suggest
*Commercial software that uses these methods was published in Nonlinear Dynamics Psychology and Life Sciences. This free version has few things missing, but many more features added, and also faster algorithms. To see my algorithm implementation for these methods, see source code.
TODO:
- I will add new tests in the future. Suggestions and algorithm codes are very welcome
- I will use Matplotlib for colored recurrence plot and other complex graphs
SEE ALSO:
Other Recurrence Plot Software.
UPDATES:
December 7, 2006:
1. After receiving a feedback from Visual Recurrence Analysis developer about the speed and recurrence plot algorithm, I decided to implement a sampling algorithm.
The exact recurrence plot option is still available in the program. If you already downloaded Bios Analyzer before, you can now just copy this new BiosAnalyzer.exe file (~134k) in the place of the old one. Others need to download the whole application (~50Mb).
November 16, 2006 - two bugs fixed:
1. For embedding plots analysis, 'Series %Isometry' and 'Series %Consecutive Isometry' titles in multiple series analysis in 'results.txt' file should instead be 'Series %Isometry - Shuffled %Isometry' and 'Series %Consecutive Isometry - Shuffled %Consecutive Isometry'. I modified this, and kept 'Series' instead of 'Shuffled' fields.
2. Regression line coefficients in 'results.txt' file were calculated with n-1 instead with n.
Download version for Windows (Linux soon to come)
Please tell me your profession and country.
DISCLAIMER:
This software is given WITHOUT ANY WARRANTY;
without even an implied statement of merchantability or fitness for a particular purpose.
If you decide to use this program, you do so entirely AT YOUR OWN RISK.
Questions from users:
Q: How to install and use it once u have downloaded, is it necessary to download python?
A: Simply unzip, and click on the BiosAnalyzer.exe file. Downloading python is not needed, as package includes relevant python libraries.
Q: What's the file format expected by this program? I have a set of data that's "time stamp(year,date,time), value" and I like to feed it through your program and see what it looks like.
A: Program assumes that only one type of data is in a file. Numbers (comma, newline, space, or tab delimited) from files are read. Non-numerical data is ignored. To read data with time stamp (year,date,time), value format, at this version of a program you have to delete columns that have date/time numbers. I will fix this for the next release (when I debug Linux version).
Screenshots of the application, and plots that it generates.
Note: recurrence plots now look differently, these are from an older version of application.
As of October 2008, Bios Analyzer users originate from more than 60 countries, several dozen professions including students, researchers, engineers, physicians, professors, psychologists, physicists, scientists, teachers, sociologists, mathematicians, etc., and very diverse institutions, some of which are: Saha Institute of Nuclear Physics, Bulgarian Academy of Sciences, The University of York, Institute of Electrical and Electronics Engineers, Purdue University, PURAC, Rush University, UNESCO Institute for Water Education, The University of Vermont, University of Warsaw, Kazakhstan Institute of Nuclear Physics, Indonesia University, Romanian Center for Complexity Studies, University of Illinois at Urbana-Champaign, Washington University in St. Louis, Indian Council of Agricultural Research.