ScatterBrain is designed to facilitate the exploration, analysis, and presentation of data sets in a hands-on, interactive fashion. In its simplest use, the points in a single x-y type scatterplot can be identified, and their underlying associated data displayed in a user-specified format in a data window. Data for multiple points can be presented, row by row, in the data window for ease of making comparisons. Individual points can be identified either with the aid of a mouse cursor, or by typing in a name associated with each row of data. In addition, multiple points can be selected and counts, totals, simple averages, and user-specified weighted averages calculated and presented in the data table.
As individual data points are identified their names can be deposited on the screen at locations determined by the mouse location. This in itself is a very useful feature when identifying points that are in close proximity to one another.
In addition, still focusing on a presentation that displays only a single graph, regression lines (also sometimes called average slope or trend lines) can be superimposed upon the plotted points and lines representing one or two standard errors of the estimate above and below the regression line may be displayed as well. Second degree polynomials can also be fitted to the data and displayed in the graphs.
One of the most unique features of ScatterBrain is that all of the options described above can be used on multiple, interrelated scatterplots. Up to four such graphs can be shown simultaneously. When a point is identified and begins to blink on and off in one graph, it will blink in the same color in all graphs. In addition, the specified data written to the data window for the corresponding row of data is color coded for ease of making a visual connection between the data shown and the location of the corresponding point in each graph. When multiple points are identified at once in one graph, all of the corresponding points are simultaneously highlighted in the same color in all of the other graphs then being shown.
Following is an example created using US Census data.
This view was created with ScatterBrain in the mode that uses multiple points. A feature permitted enclosing approximately 10-11 percent of all students in each of the two sets of districts selected. The scatter plot presents districts based upon the percentages of heads of households who did not have high school diplomas (y-axis) versus the percentages of heads of households having college degrees (x-axis). The two rows of data show that although both sets of districts spent approximately the same amount per pupil, those with less-educated parents had nearly 7 times the poverty rate among children, one-third the household income, and lived in homes less than a third of the value of those housing families in the more educated group. There is a distinct geographical pattern to these points, which is quite visible.
ScatterBrain is written in Java. Consequently, the executable file that is downloaded when you click on the link below includes a Java run-time module.
The program is bundled with the data used in creating the last two exhibits above. Ordinarily, to get from raw data to a presentable chart requires several steps--choosing variables to be plotted, establishing scales, reference lines, and so forth. The package to be downloaded contains setup files that create several graphs, formats the associated data tables, and instantly puts them on the screen. They are closely tied to exhibits that appear in the preliminary paper referenced, and linked-to, above.
If you choose, you can modify the graphs and
options and save those changes in a new setup file, and save it on your own storage
device. Once created, either manually from scratch or by a previously
saved setup file, the graphs can be interactively queried.
The complete User Guide is also available for download below. It is in Microsoft Word (c) format. The User Guide really must be read and studied if you choose to try the program.
Click to download:
ScatterBrain User Guide
ScatterBrain with Texas Big90 data
The version of
ScatterBrain included in these download files is fully functional. The data
files used are simply tab-delimited text files with field names in the first
row, integer row ID numbers in the first column, alphanumeric names associated
with each row in the second column, and alphanumeric names, beginning with
letters A..Z, for any categorical variables. All other columns contain
My son, Laurence P Toenjes, made substantial contributions to this project.
If you would like to try ScatterBrain with your own files send me an email. This offer is only for non-commercial use.
711 West Shore Drive,
Clear Lake Shores, TX 77565