Summary

Specialisation Data Analysis and Interpretation
Course Data Management and Visualisation
Education Institution Wesleyan University
Publisher Coursera
Assignment Running Your Third Program

Exploratory Data Analysis

Source

The Mars Craters data set was made available by Wesleyan University/Coursera as part of the Data Management and Visualisation course, of the Data Analysis and Interpretation Specialisation, from the Ph.D. Thesis Planetary Surface Properties, Cratering Physics, and the Volcanic History of Mars from a New Global Martian Crater Database (2011) by Robbins, S.J., University of Colorado at Boulder.

Size

The data set has a total of 384343 observations and 10 variables.

The variables are: CRATER_ID, CRATER_NAME, LATITUDE_CIRCLE_IMAGE, LONGITUDE_CIRCLE_IMAGE, DIAM_CIRCLE_IMAGE, DEPTH_RIMFLOOR_TOPOG, MORPHOLOGY_EJECTA_1, MORPHOLOGY_EJECTA_2, MORPHOLOGY_EJECTA_3 and NUMBER_LAYERS.

Univariate Analysis

Hemisphere

Hemisphere is a variable derived from the LATITUDE_CIRCLE_IMAGE variable to transform the continuous coordinates into categories, for the sake of brevity.

Hemisphere shows seven occurrences in the Equator, same as Latitude equals to zero. Just above 60% of the observations are located in the South Hemisphere. Also, all the observations have values.

Ejecta Morphology 1 (Group by Main Feature)

The variable MORPHOLOGY_EJECTA_1 has 339718 out of 384343 values missing, or 88.3%. The recording with existing content are divided in a large number of categories if considered the full morphology qualification. If taken into account just the first classification, the number of categories is reduced to 29.

From the recorded data, considering just the first classification, shows that 27069, or 60.6%, are of the RD category. The only two other categories that have more than 10% are SLERS (11.45% = 5111) and SLEPS (11.20% = 4998).

Discussion

The morphology of the craters could lead to analysis of particular events concentrated on either hemispheres on Mars.

From this cross tabulation and charting of Hemisphere and Morphology Eject 1 it is possible to verify that Morphologies such DLEPCPD, DLEPD, DLEPSPD and SLERCPD are only present in the North Hemisphere, while DLERSRD, DLSPC, MLERSRD, SLEPCRD, SLEPSPD and SLERSRD are only present inthe South Hemisphere.

SAS Code

/* Use Course's Library */
LIBNAME mydata "/courses/d1406ae5ba27fe300" ACCESS = readonly;

/* Configure the Data */
DATA NEW;
  /* Data set */
  SET   mydata.marscrater_pds;
  LABEL Hemisphere    = "Hemisphere"
        MorphoE1U     = "Ejecta Morphology 1 (Group by Main Feature)";

  /* Categorise the Latitude in Hemispheres */
  IF (LATITUDE_CIRCLE_IMAGE < 0)
    THEN Hemisphere = 0; /* South */
    ELSE Hemisphere = 1; /* North */

  /* Collapse the Morphology of Eject 1 to its Main Feature, to reduce the output */
  IF (INDEX(MORPHOLOGY_EJECTA_1, "/") = 0)
    THEN MorphoE1 = MORPHOLOGY_EJECTA_1;
    ELSE MorphoE1 = SUBSTR(MORPHOLOGY_EJECTA_1, 1, INDEX(MORPHOLOGY_EJECTA_1, "/") - 1);
  MorphoE1U = UPCASE(TRIM(MorphoE1));

PROC SORT;
  BY CRATER_ID;
RUN;
PROC GCHART;
  HBAR Hemisphere/DISCRETE TYPE=PERCENT;
PROC GCHART;
  HBAR MorphoE1U/DISCRETE TYPE=PERCENT;
PROC GCHART;
  VBAR MorphoE1U/DISCRETE TYPE=MEAN SUMVAR=Hemisphere;