Vous êtes sur la page 1sur 3

Taylor J Dilliott, BA

1-109 Sanford Ave South, Hamilton, ON L8M 2G7


E: Taylor.J.Dilliott@gmail.com M: (289) 969-1847

May 27, 2018


Mr. Ian D. Smith
Campus Facilities Management Supervisor
Niagara College, Niagara-on-the-Lake Campus
135 Taylor Road, Niagara-on-the-Lake, ON, L0S 1J0

Dear Mr. Smith

RE: GISC9318- Spatial Statistics D1

Please accept this email as a formal letter of transmittal for GISC9318 D1- Introduction to
Statistics Mapping as per the terms of reference. Attached to this email you will find the Excel
Worksheet, a map document displaying the well locations from 2015 and the histogram and
boxplot that were created in R Studio.

The first set of data from 2015 was used to determine the following statistics:

1. The water table depths range from -0.1 meters to 21.4 meters from the top of the well
casing. The -0.1 meter depth can be explained by the water being above ground level,
in this case based on the map created using XY data it appears the well is located near
the campus lagoons.
2. The pH value (Log H*) of the water wells on campus ranges from 4.6 to 8.5.
3. The average water table depth is 8.6 meters while the median depth is only 6.5 meters.
This variance is caused by the extremely deep wells located on the escarpment skewing
the data.
4. The histogram created in Excel using the 2015 data is skewed to the left, suggesting that
the average is higher than the median, which was shown in question 3.
5. The two wells located on the escarpment could be considered outliers as they are
significantly deeper than any other wells and have a lower pH than any of the other
wells. However, given that they follow the negative correlation shown by the other wells
when placed on a scatter plot it is more likely that they would be considered normal
significant values in a larger dataset. In reality there is always a potential for outliers
to exist depending on external factors such as time of year, landscape conditions and
human error.
6. Given the water table depths standard deviation of +/- 7.1 meters, it could be
reasonably expected that 68% of the 100 new wells would fall somewhere between 1.5
and 15.7 meters in depth.
7. Using the calculated standard deviation for the water table depths of +/- 7.1 meters it
could be assumed that 95% of the 100 new groundwater wells would have a depth ranging
from -5.6 meters to 22.8m.
GISC9118- Foundations of Mapping and Cartography D3
Gradient Mapping
May 27, 2018
8. Based on the scattergram created from the 2015 data, where water table depth is the
independent variable and the pH (Log H*) value is the dependent variable there is a
strong linear correlation between the two variables.
9. Continuing to work with the scattergram, the relationship between groundwater depth
and pH level has a negative correlation, showing that the deeper the well is, the lower
the pH is likely to be.
10. Finally, using the scattergram the relationship between the water table depth and pH
level could be described as strongly linear. This is based off the correlation coefficient
(Pearson r) having a value of -0.9464 which shows a very strong negative correlation
between the two variables.

The 2016 dataset “D1 – New Mini-Piezometer Data” was loaded into R Studio and analyzed to
answer the following:

11. With the new dataset and measurements the mean water table depth changes from 8.6
meters to 8.4 meters.
12. The standard deviation of the well depth also changed from 7.1 meters to 6.7 meters.
This shows that the well depths are more clustered with fewer outliers on the high and
low ends.
13. The histogram is attached to this email as 2016DepthHistogram.bmp
14. Based on the boxplot created in R Studio (Attached to this email as
WaterTableDepthBoxplot.bmp) there are 5 potential outliers represented by the small
circles on the plot. These can be classified as potential outliers because they fall well
outside the range of the upper quartile.
15. A) Using a simple linear regression model provides an equation of pH=7.88618 + Depth *
(-0.167), where Depth is the independent variable, 7.88618 is the intercept coefficient
and -0.167 is the regression constant. The R2 value provided by the model is 0.8983,
which shows a strong linear correlation. The adjusted R2 value is 0.8856, which still
shows a strong linear correlation between the two variables.

B) Creating a multiple regression model in R using pH as the dependent variable and


Depth, Easting and Northing as independent variables yields an equation of pH= 5497 +
Depth(--0.2123)+ Easting(0.001) + Northing(-0.001). The R2 value for this model is 0.9697
which is a fairly significant increase over the simple regression models R2 value of
0.8983. The adjusted R2 value is 0.9545, which again is a significantly higher value than
the simple regression model and shows a very strong correlation between the pH, depth
of the well and its geographic location. Based on the model, none of the values could
be considered truly insignificant, though the Easting variable is the least significant.

C) While both models would do an adequate job predicting the pH values, the multiple
regression model that used all the variables will likely yield the best predictions for pH
based on the adjusted R2 value of 0.9545.

1-109 Sanford Ave South, Hamilton, ON L8M 2G7


E: Taylor.J.Dilliott@gmail.com M: (289) 969-1847
Page | 1
GISC9118- Foundations of Mapping and Cartography D3
Gradient Mapping
May 27, 2018
D) For the wells data none of the independent variables present a strong correlation,
with only depth and Easting presenting a moderate correlation (0.6733). Due to this no
adjustments need to be made in order to avoid multi-colinearity.

I found this assignment to be an excellent introduction to working with R and a limited dataset
and look forward to applying these skills to a larger dataset in the later assignments. I
particularly appreciated how the assignment recalled examples given in class and in the
lectures.

If you have any concerns or issues with the answers or any of the attached documents please
do not hesitate to contact me.

Sincerely,

Taylor J. Dilliott, BA
GIS-GM Certificate Candidate

TJD\

Enclosures: DilliottTGISC9318GroundwaterData.xlsx
WellLocationMap.JPG
2016DepthHistogram.bmp
WaterTableDepthBoxplot.bmp

1-109 Sanford Ave South, Hamilton, ON L8M 2G7


E: Taylor.J.Dilliott@gmail.com M: (289) 969-1847
Page | 2

Vous aimerez peut-être aussi