Global Data Competition Submission

  I have been busy last week working on the analysis of climate data and air quality data for Utah in 2013 for the Global Data Competition.  Above is the document I submitted and my team put together in this graphical display (Thank you Jim, Nikki, Jackson, and Nimesh),  

The first page in the embed seen above give the descriptive statistics for the analysis.  The next one is a mixed model analysis of the merged data set for three air pollutants (Carbon Monoxide, Nitrogen Dioxide, and Sulfur Dioxide) stations and the outcome of departure from normal temperature (DPNT) in tenths of degrees Fahrenheit.  We found no effect of pollutants but there was an effect of elevation with higher levels elevation in meters and a significant random effect of date.  Date was by month.

The model was rerun with the full data set.  There was a significant 0.0004 degree increase in the departure from normal temps for every 1 meter gain in elevation.  There was a significant random effect of date.

Jnnuary correlation

Because of this result I looked at the correlation with DPNT and elevation for each month in 2013 in Utah.  The correlations are presented in the table at the bottom of this post.  There were significant positive correlations for the winter months of December thru March with DPNT increasing in each of these months with elevation. The scatter plot for January where the relationship was strongest. In the summer months of July thru September, there were significant negative correlations which were weaker in magnitude (distance from zero) than those for January and December.  The scatterplot for July (the strongest negative correlation) is presented below.

July Correlation

The significance of higher temperatures at higher elevations in the winter months means a smaller snow pack in the mountains.  This has implications for the ski industry in Utah as well as the water tables in Utah which depends on the spring snow melt.  This can contribute to drought conditions throughout the west.  The winner of the competition will be announced at the Intermountain Data Conference on November 21.

Month	Correlation	P-value
January	0.4507225	0.0001
February	0.2918966	0.0021
March	0.2167232	0.0236
April	-0.1209575	0.2102
May	-0.1022036	0.2926
June	0.0381272	0.6925
July	-0.3159874	0.0009
August	-0.2779292	0.0031
September	-0.2903617	0.0021
October	-0.1866295	0.0578
November	-0.0167705	0.8652
December	0.4267564	0.0001

**Related Posts**

What Shaq can teach us about Climate Change

Global Warming, Wikileaks, and Statistics: What Barry Sanders Can Teach Us.

The Impact of the ACA (Obamacare) in Utah and Pennsylvania by Ethnicity and Race