Assignment 1_RMIP: Writing About Your Data

This blog will cover the assignments for the Regression Modelling in Practice course. It is a continuation of the Data Analysis and Interpretation Specialization taught by Wesleyan University under Coursera training platform.

The week 1 assignment is about describing my data; in terms of the population my sample was drawn from, the procedures that were used to collect the data, and the measures that I'll use in my statistical analyses.

 

Sample

The sample is from a new global database of Mars Impact Craters created by Stuart Robbins as part of his PhD thesis (http://about.sjrdesign.net/research_mars.html) carried out between 2008 and 2011. It contains individual properties and parameters of 384, 343 Martian impact craters. Stuart’s global Mars crater database is extensive because it contains complete set of information for large craters and also smaller craters with diameter less than 1km. The global Mars database contains craters with measured crater depth ranging from -0.42km to 4.95km. Crater diameter is between 1km and 1164.22km. Crater latitude and longitude were measured in decimal degrees North and decimal degrees East respectively. Recorded number of crater ejecta layers is from 0 to 5. Craters were mainly classified into various morphology groups; mainly simple and complex morphology categories.

The data analytic sample for this study includes only 18, 065 fresh non-eroded craters. The criteria for choosing this sample subset include craters with depth greater than 0 and less than or equal to 3km; craters with diameter greater than 0 and less than or equal to 100km and craters with number of ejecta layers greater than 0.

Procedure

From satellite images and existing databases, Stuart used modern digital techniques and algorithms to identify craters and determine their fundamental properties such as depth, diameter, location (latitude & longitude), morphology and number of ejecta layers. The crater identification and classification procedure was done mainly in ArcGIS software using infrared mosaic observations captured by a Thermal Emission Imaging System (THEMIS) aboard the 2001 Mars Odyssey NASA spacecraft.

Measures

My main research question is to determine if crater depth is associated with crater diameter and to verify whether this association is dependent on crater location on Mars (crater latitude). Crater depth is the response variable while crater diameter is the explanatory variable. Crater latitude will be used as a moderator or as a confounding variable.

Crater depth is a quantitative variable which measures the average elevation from the depth floor to the depth rim (units are in kilometres). Crater diameter is also a quantitative variable which measures the size of craters from a derived centre of a non-linear least-squares circle fit (units are in km). Location was measured by two quantitative variables – crater latitude and crater longitude. For the current analysis, crater latitude was binned into four categories; each containing 45 degrees bin interval from the North Pole to the South Pole. This new collapsed variable is called MARS_REGION. Crater depth and diameter will be used as quantitative variables during the regression analysis.

 

Author:
Posted on February 07, 2016 by Okechukwu Ossai

Posted in Regression Modeling in Practice Course.

admin

Leave a Reply

Your email address will not be published. Required fields are marked *