* Figure 1.1 from my Master of Sciences thesis
Although you may not always think about statistics when thinking about graphic design, graphics are extremely important in statistics. Statistics software provides ways to illustrate data interpretation.
To generate these graphics, a working knowledge of the calculations behind the statistics is important. For example, inappropriate uses of ordination techniques will generate figures which are difficult to ascertain anything useful from.
I have taken three graduate level courses in statistics at the University of Tennessee. My software experience includes SAS and R. My experience has concentrated in linear mixed models in SAS and R for ordination, QTL mapping, and other genetic analysis.
Graduate level statistics courses
| Course |
Credit hours |
Grade |
| Statistical Genomics |
3 |
A |
| Least Squares Analysis |
3 |
A |
| Statistics for Biological Research |
3 |
A |
* Figure 1.4 from my M.S. thesis.
Here are two works where I was required to take raw data tables and develop figures that interpret the data.
*Figure text
Figure 1.1: Bee visitation on one East Tennessee farm represented by nonmetric multidimensional scaling (NMDS) ordination. Unlabeled points represent transect observations. Labeled crosses represent bee category scores. Ellipses are the class factor of flower type and 95% confidence intervals of the standard deviation of their weighted averages. Permutation testing of flower type within day shows this factor is significant (P < 0.001). A vector fitted to the time of day within flower type is significant (P < 0.001). Thin plate regression splines further show the effect of time of day on the ordination using a generalized additive model with family = quasipoisson.
Figure 1.4: Bee visitation to squash and pumpkin across 10 East Tennessee farms represented by NMDS ordination, with environmental vectors and regression splines for time of day. Unlabeled points represent transect observations. Permutation testing within transect specified as strata shows the vectors are significant for date (P < 0.001), time of day (P = 0.018), and female flower ratio (P < 0.001).
** Thesis abstract: Wilson, Michael Edward, “Bee Visitation to Crops and other Flowers Planted as Bee Food. ” Master’s Thesis, University of Tennessee, 2011. http://trace.tennessee.edu/utk_gradthes/1109
Farms that require insect pollination and reside in diverse landscapes benefit from pollination by native bees. However, populations of native bees and honey bees are generally in decline and this threatens food production. Documentation of crop pollination as an ecosystem service is needed to identify potential impacts from declining bee populations. This study identifies communities of bees providing pollination and how they vary across different crops and environmental conditions. Managing landscapes to provide additional food sources for bees may improve the health of wild and managed bees. This study also evaluated the attractiveness of bees to selected species of plants that could be used to provide food sources.
In 2008-09, bee visitation was measured on 10 different crops among 12 farms in Tennessee. On one of these farms, visitation was observed for 24 different flower species that could be used for supplemental bee forage. Bees visiting flowers were organized within a classification scheme of 10 taxonomic groups. Environmental data for each observation was recorded including the type of flower, the date, time, location, farm and plot size, if the farm was organic or conventional, and the number and sex of flowers when appropriate.
Nonmetric multidimensional scaling and linear models revealed that native bees are important visitors to crop flowers, but their abundance and composition depend on the type of flower. Within a flower type, other environmental effects can shift the community composition. Plants selected for habitat enhancement can be chosen based on the similarity of the community of bees which utilize them as compared to crops.
