Will my data and functions work with Rave?

From Rave Documentation
Revision as of 15:58, 22 August 2013 by Admin (Talk | contribs) (Data Set File Formats)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Data Set File Formats

Rave supports "flat file" type data sets, i.e. data that can be represented as a table in which each column contains a variable and each row contains a set of observations of each variable (i.e. a "data point"). The data may be numerical, text, or a mix of both (although all entries within a column must be the same type).

Rave can import data from Excel files (.xls or .xlsx), tab-delimited text files, or comma-delimited text files (.txt or .csv).

For more details about how the data files must be formatted, see: Data Set File Formats

Data Set Size Limits

The size of data you can work with in Rave is limited by your computer's RAM. As a rule of thumb, the size of the data you are working with should be less than 1/8th of the amount of RAM you have. This is because Rave stores multiple copies of your data (a working copy and a backup copy), as well as some metadata whose size scales with the size of your data.

When working with large data sets (e.g. more than a few thousand data points), you may want to avoid certain visualizations whose complexity scales with the size of your data set. You may notice a significant slow-down when using such visualizations with large data sets. If you simply want to create a visualization and export it to a file, the slowdown might be OK, but if you are creating an interactive dashboard, it will probably be unacceptable.

Things to avoid when working with large data sets include:

  • Scatter plot matrices
  • Scatter plots that use the "Color points by..." (except color by Row Colors) or "Size points by..." options. (These options require Rave to use more complicated plot objects that are significantly slower)
  • Parallel coordinates plots (Although you may use the "Do not draw more than X lines" option on the View Tab to avoid a slow-down)


Function Requirements

Rave can work with just about any data-generating MATLAB function, although it requires a particular format for the inputs/ouputs of the function and the function definition line. It also requires that the function be able to handle evaluating multiple data points in a batch.

This may require you to write a new "wrapper" function that acts as an interface between Rave and your existing MATLAB functions. For more details, see User-Supplied Matlab Function Formats