INTRODUCTION
ARGUS LITE is a nifty little system that will let you enter data in Matrix format (Rows are OBJECTS; Columns are VARIABLES), and will then generate a Perceptual Map from this DATA.
Perceptual Mapping is USEFUL because it allows you to VISUALIZE numerical data in CONDENSED form. It's essentially a DATA REDUCTION and/or PATTERN RECOGNITION tool. A Matrix with 10 columns would require 45 two dimensional SCATTERPLOTS to show all UNIQUE 2-WAY combinations of OBJECTS and VARIABLES, while ARGUS LITE can show the same information on a SINGLE 2-D map.
ARGUS LITE will help you to see PATTERNS in NUMBERS stored as TABLES, and will also tell you which OBJECTS are similar/dissimilar, which VARIABLES are measuring the same underlying characteristic, and which OBJECTS are most closely associated with which VARIABLES.
You need MINIMAL HARDWARE to run this system (DOS 3.0 or up; 200k disk space; 200k free memory; ASCII graphics; OPTIONAL printer). It's comfortable to use (MENU/PROMPT driven with ON-LINE documentation) and requires very few decisions to provide USEFUL results.
DESIGN PHILOSOPHY
ARGUS LITE is a STREAMLINED (hence LITE in the name) system for producing Perceptual Maps. Data entry is accomplished via a simple TEXT EDITOR and you don't have to specify any options to generate a map from the data. We have tried to make an inherently complex operation as simple and intuitive as possible by limiting the number of choices and keystrokes (EASE of USE = Visible Options + Few Operations).
Everything is MENU/PROMPT driven (POINT n' SHOOT), and the only TYPING you have to do is when you enter your own data.
The basic unit of analysis is an AGGREGATED Data Matrix. This could be derived from a much larger Data Base. In fact, it's very likely that you will get the Matrix INPUT from SECONDARY data, another DATA BASE, or a STATISTICAL ANALYSIS program. A CROSSTAB or a TABLE consisting of AVERAGE RATINGS of OBJECTS on ATTRIBUTES represents typical INPUT for analysis. As a result, the capacity restriction of ROWS x COLUMNS <= 320 is not that serious, and essentially represents the limits of what you can handle visually in a single 2-D map.
The mechanics of program implementation are designed to SAVE SPACE. Everything is COMPRESSED. Specific components are extracted as needed, and REMOVED after they've been used. The program CLEANS UP AFTER ITSELF.
While this tends to SLOW things down a bit and presents you with a lot of meanigless screen display (just ignore it), you'll need a workspace of less than 200K to accomodate elements that would require in excess of 400k if fully expanded (IDEAL for antique HARDWARE and LAPTOPS).
The 'LITE' nature of the program has some limitations (SLOW EXECUTION). It's designed to give you only a TASTE of what Perceptual Mapping can do.
If you hunger for MORE, consult the REGISTER/UPGRADE section of this document. ARGUS LITE and all its components are absolutely FREE. You don't have to register ARGUS LITE. A REGISTRATION/UPGRADE consists of ordering an entirely DIFFERENT product with similar functionality but enhanced capability.
ENTERING/EDITING Data Matrix
To enter your own data for mapping, select [ENTER/EDIT Data Matrix] from the menu. You'll see a brief instructional screen, and if you decide to proceed, a PROTOTYPE data matrix will be preloaded in the editor. It looks like this:
C:mydata.asc Line:00 Col:05 Free:6138 Mode: EDIT Com key:[F1] --- TITLE --- --- col001 col002 col003 --- --- ROW001 000.00 000.00 000.00 --- ROW002 000.00 000.00 000.00 --- ROW003 000.00 000.00 000.00 --- ROW004 000.00 000.00 000.00 --- ROW005 000.00 000.00 000.00 ---
To enter your own matrix, edit over the prototype. However, before you do this, keep in mind that there are some format conventions that must be observed.
1) ALL non-numeric entries (TITLE; LABELS) must start with an
ALPHABETIC character.
2) The TITLE must start at the LEFT margin with no leading
spaces (FLUSH LEFT). It can consist of multiple words
with a total length of less than 20 characters.
3) The ROW LABELS must also start at the LEFT margin with no
leading spaces (FLUSH LEFT). ROW LABELS can contain NO spaces
and must be EXACTLY six characters long. Row labels should
be in UPPER CASE. This helps to identify ROWS (Objects) on
the map.
4) The COLUMN LABELS must be preceded/separated by at least ONE
BLANK SPACE. COLUMN LABELS can contain NO spaces and must be
EXACTLY six characters long. Column labels should be in
lower case. This helps to identify COLUMNS (Variables) on
the map.
5) Each NUMERIC ENTRY must be preceded/separated by a BLANK SPACE.
You don't necessarily have to have a decimal point in a numeric
entry.
For MISSING data you can either enter/leave a ZERO cell (bad idea)
or try to estimate a COLUMN MEAN and enter it.
6) In general, LABELS and NUMBERS are delimited by SPACES (so
don't use SPACES in LABELS), and each LINE is ended/delimited
by a LINEFEED/CARRIAGE RETURN.
7) The CAPACITY of the system is limited to ROWS x COLS < 320.
The number of COLUMNS cannot exceed the number of ROWS.
Thus with a limit of ROWS x COLS < 320, the MAXIMUM number
of columns is 17.
You're least likely to have problems if you edit directly over the labels and numbers in the PROTOTYPE. The EDITOR is in OVERWRITE mode, so if you type over something, what you type will replace what's there.
Use the ARROW KEYS to move around. To add ROWS/COLUMNS just type them in. Keep to the same format as the PROTOTYPE. To delete ROWS/COLUMNS, just put the cursor over what you want to remove, and hit the [Del] key.
When you've finished entering your matrix, hit [F1] and then [A]. Your matrix will be saved in a file named MYDATA.ASC. Type in this file name when the prompt appears. The file MYDATA.ASC will now be converted to MYDATA.DPM. If you then want to produce a map of your own data, MYDATA.DPM is the file you should select.
GENERATING a Perceptual Map
To generate a map, first select [GENERATE Perceptual Map] from the menu. You'll see a brief instructional screen, and if you decide to proceed, a file selector screen will appear. The screen looks like this:
copy # aplinput.dpm MUFFJS.DPM 640 15.11.85 00:03 MYDATA.DPM 166 2.03.95 17:14 TURBO.DPM 225 11.02.95 17:00 ->AFCTOT.DPM 861 19.07.94 10:10
Let's assume you want to select AFCTOT.DPM. Move the cursor with the DOWN ARROW until AFCTOT.DPM is highlighted. Next, press the [SpaceBar] to MARK the file. An ARROW will appear to the LEFT of your selection. Finally, press [Enter] to CONFIRM.
The selector screen will disappear, and after a lot of on-screen activity which you CAN IGNORE, a MAP will eventually be produced. This process can be TIME CONSUMING. If nothing seems to be happening, it's not that the program is HUNG UP. It's just S-L-O-W because it's INTERPRETED, so be patient and have a cup of coffee while it's doing its work.
When the MAP is finally in view, a [ YES ] NO prompt at the bottom left of the screen will ask if you want to 'Show labels'. Press [Enter] for 'YES', and press [Enter] again when you're done looking at the map.
READING RESULTS
If you yourself had entered the file that was selected in the example (AFCTOT.DPM), your INPUT would look as follows:
AFC OFFENSE '83
**Wins TD-Tot PtsTot YDPass TDPass YD-Run TD-Run *X-Pts FldGls OppPts
PITTSB 10.0 39.0 355.0 2754.0 15.0 2610.0 17.0 38.0 27.0 303.0
CINCIN 7.0 43.0 346.0 3492.0 14.0 2104.0 24.0 40.0 16.0 302.0
NEWENG 8.0 36.0 274.0 3040.0 16.0 2605.0 19.0 31.0 9.0 289.0
HOUSTN 2.0 34.0 288.0 3286.0 16.0 1998.0 16.0 33.0 17.0 460.0
BALTMR 7.0 25.0 264.0 2663.0 12.0 2695.0 10.0 22.0 30.0 354.0
*MIAMI 12.0 48.0 389.0 3235.0 28.0 2150.0 16.0 45.0 18.0 250.0
RAIDRS 15.0 54.0 442.0 3910.0 31.0 2240.0 18.0 51.0 21.0 338.0
SEATLE 11.0 50.0 403.0 3316.0 25.0 2119.0 19.0 49.0 18.0 397.0
KANCIT 6.0 45.0 386.0 4684.0 29.0 1254.0 13.0 44.0 24.0 367.0
SANDGO 6.0 45.0 358.0 4891.0 27.0 1536.0 16.0 43.0 15.0 462.0
BUFFLO 8.0 36.0 283.0 3438.0 30.0 1736.0 4.0 34.0 11.0 351.0
CLVLND 9.0 42.0 356.0 3932.0 27.0 1922.0 13.0 38.0 22.0 342.0
NYJETS 7.0 38.0 313.0 3742.0 21.0 2068.0 11.0 37.0 16.0 331.0
DENVER 9.0 34.0 302.0 3466.0 17.0 1784.0 15.0 33.0 21.0 327.0
This Data Matrix represents team football statistics for the AFC conference
of the NFL for the 1983/4 season. The ROWS are the teams and the COLUMNS are
the various performance statistics. Once this input has passed through the
computational process, a map is produced that looks like the one below:
AFC OFFENSE 83'
+-------------------------------------------------------------------+ 1.0
| *X-PtsTPtsTot |
| | |
| TDPass | |
| YDPass | |
| | | 0.6
| | **Wins |
| | RAIDRS |
Y | | |
C | KANCIT | |
O |------------SANDGO-----------SEATLE--*MIAMI--TD-Run--------------| 0.2
O | | |
R | OppPts CLVLND |
D | CINCIN |
| BUFFLO NYJETSFldGls |
| DENVER PITTSB | -0.2
| HOUSTN | |
| NEWENG |
| | |
| BALTMR YD-Run |
+-------------------------------------------------------------------+ -0.5
-0.69 -0.31 0.06 0.44 0.81
XCOORD
There are a number of features in this map. First of all, the POINT locations on the map are represented by the ' ' symbol. The LABELS for each POINT are either to the RIGHT or the LEFT of the POINT depending on whether the POINT is in the LEFT or RIGHT side of the graph.
All the OBJECTS (in this case TEAMS) have labels that are ALL CAPS, while the VARIABLES are in UPPER/lower case. This helps in recognizing each.
The VARIABLES tend to be nearer to the edges of the map, while the OBJECTS tend to be closer to the center.
VARIABLES
VARIABLES that are measuring the SAME underlying characteristic, tend to be close together. Thus (in this example) variables related to SCORING (i.e. X-Pts [Extra Points]; TD-Tot [Total Touchdowns]; and PtsTot [Total Points]) are all close together or OVERLAPPING on the map. VARIABLES such as YDPass [Yards Passing] and YD-Run [Yards Running] which are measuring OPPOSITE characteristics are FAR APART on the map. VARIABLES which are essentially INDEPENDENT of each other such as TDPass [Touchdowns Passing] and **Wins [Wins] are at RIGHT angles from each other.
OBJECTS
OBJCTS that tend to have SIMILAR profiles, are close together. Thus SANDGO and KANCIT [San Diego Chargers and Kansas City Chiefs] are close together because they both passed a lot and had (in 83/4) weak defenses. Conversely, these OBJECTS are far from BALTMR ( now the Indianapolis Colts) because that team was strong on running and defense back then. Neither SNDGO/KANCIT nor BALTMR was high on **Wins, which more balanced teams such as the RAIDRS, MIAMI, and SEATLE were relatively higher on.
OBJECTS on VARIABLES
In general, the closer an OBJECT is to a VARIABLE on the map, the more associated it is with the characteristics represented by the VARIABLE.
REGISTERING/UPGRADING
ARGUS LITE is FREE.
If you'd like to examine a very small sampling of the GRAPHICS available in each of our software products, JUMP to our GRAPHICS SAMPLER.
If you want to try ARGUS LITE 'via DOWNLOAD', by all means link to our: FREE DEMO SOFTWARE.
o Return to: 'SPRING SYSTEMS HOME PAGE'