1 .R Packages required

Make sure you have latest R and Rstudio installed before starting this process. These are the R packages that are required to complete the data cleaning and documentation using Rstudio.

knitr - for rendering HTML reports
tidyverse - for data manupulations

Note: The above packages do not come with Rstudio installation, they need to be installed explictly, use the packages tab or just type install.packages(“package_name”).

Next load the R packages:

#install those packages before run the rmd if there is no package following codes:
#install.packages(c("rio","data.table","dplyr","anytime","knitr","kableExtra", "ggplot2","ISOweek","alluvial","lubridate"))
library("knitr")
library("tidyverse")

2 .Avian Monitoring

2.1 Data Preparation

Reshaped data of Avia Monitoring

Monitoring_Route	Date	eBird	eachCount	totalCount	H_S	variable	value	Weeks	WeekNoYear	Year	Month	MonthWithYear	season
Blue Trail	2018-06-27	N	14	10	Seen	GBHE	2	2018-W26	W26	2018	6	2018-06	summer
Blue Trail	2016-07-13	N	28	14	Heard	MAWR	1	2016-W28	W28	2016	7	2016-07	summer
Purple Trail	2017-08-28	N	11	6				2017-W35	W35	2017	8	2017-08	summer
Blue Trail	2016-07-30	N	15	5	Seen	GBHE	4	2016-W30	W30	2016	7	2016-07	summer
Blue Trail	2017-03-13	N	12	4	Seen	CAGO	8	2017-W11	W11	2017	3	2017-03	spring

Aggregating by each species and week with no year

WeekNoYear	variable	value
W14	GBHE	1
W22	WODU	4
W11	CAGO	11
W26	GBHE	7
W22	WOTH	3

Data from eBird

Week_starting_on	species	Frequency	Total_checklists_submitted	Abundance	Birds_Per_Party_Hour	checklists_reporting_species	High_Count	checklists_reporting_species__1	Totals	checklists_reporting_species__2	Average_Count	checklists_reporting_species__3	variable
01-07	Canada Goose	37.442681	5670	50.3238095	385.820322	2137	4500	2293	304769	2187	139.354824	2187	CAGO
10-21	Great Blue Heron	23.639091	4666	0.5816545	4.145696	1128	74	1144	2820	1137	2.480211	1137	GBHE
07-21	Wood Thrush	9.742288	5122	0.1805935	2.445683	511	12	539	971	527	1.842505	527	WOTH
07-31	Wood Duck	16.662159	3697	1.6418718	10.848080	618	86	624	6171	618	9.985437	618	WODU
02-07	Wood Thrush	0.000000	7679	0.0000000	0.000000	0	0	0	0	0	0.000000	0	WOTH

2.2 Descriptive Analysis

2.2.1 Species Number in Sample

Data Summary:

This plot shows the total number of each species based on the csv file. As per the plot we can see that CAGO is the highest in number and WOTH is lowest in number. This refers to that CAGO are the most seen species and GBHE is the second highest in number. Also, the data shows that MAWR and WODU are almost similar in number.

2.2.2 Alluvial Diagram

Data Summary:

a, CAGO(Canada Goose) and GBHE(Great Blue Hero) are the same magnitude, MAWR(Marsh Wren), WOTH(Wood Thrush) and WODU(Wood Duck) are another size group in this sample.

b,2017 has the largest sample number, but 2016 and 2018 are not much less than 2017. Different species have different composition ratio.

c, The way of observation in this five species concentrates on Seen. MAWR and WOTH have a large proportion on Heard than Seen. MAWR has no sample on H&S.

d, Most of the proportion in the sample comes from Blue Trail, but most WOTH number comes from Red Trail.

e, Each species has its unique season to be observated.

CAGO appears in winter,spring and autumn.
MAWR only appears on summer.
GBHE is mostly observated in summer,and part is in spring.
Most of WODU are in summer, and a little of them are in autumn.
WOTH is only in summer.

2.2.3 Cross Table

With season

variable	autumn	spring	summer	winter
CAGO	22	63	1	31
MAWR	0	1	19	0
GBHE	1	18	85	0
WODU	5	3	16	0
WOTH	0	0	14	0

With Monitoring Route

variable	Blue Trail	Green Trail	Purple Trail	Red Trail
CAGO	105	6	2	4
MAWR	20	0	0	0
GBHE	103	1	0	0
WODU	24	0	0	0
WOTH	0	1	1	12

With H_S

variable	H&S	Heard	Seen
CAGO	2	12	103
MAWR	0	14	6
GBHE	10	10	84
WODU	6	0	18
WOTH	1	10	3

With Year

variable	2016	2017	2018
CAGO	24	72	21
MAWR	11	1	8
GBHE	30	33	41
WODU	13	9	2
WOTH	7	2	5

2.3 ANOVA Analysis

According to descriptive part, We can be sure each variable with each species in the sample is not homogeneous. But five species have their unique characters on each variable, and we may eplore more relationship on the variables.

The total number of observation in MAWR, WOTH and WODU is not much enough to well estimate the variables difference in the sample. We will only take a trying to test anova group defference on the two species(CAGO, GBHE).

2.3.1 Canada Goose(CAGO)

#Randomized Block Design
fit <- aov(value ~ H_S + Year + Monitoring_Route + season, aviaAllu[variable == "CAGO",])
anova(fit)

## Analysis of Variance Table
## 
## Response: value
##                  Df Sum Sq Mean Sq F value Pr(>F)
## H_S               2 565.79 282.893  2.0096 0.2488
## Year              1   7.49   7.494  0.0532 0.8288
## Monitoring_Route  3 301.97 100.657  0.7150 0.5924
## season            3  86.87  28.956  0.2057 0.8876
## Residuals         4 563.10 140.774

As the result shows, we can not reject the null hypothesis on 5%, GAGO’s observation number in the four variables has no significant different.

2.3.2 Great blue hero(GBHE)

#Randomized Block Design
fit <- aov(value ~ H_S +  season + Year + Monitoring_Route, aviaAllu[variable == "GBHE",])
anova(fit)

## Analysis of Variance Table
## 
## Response: value
##                  Df Sum Sq Mean Sq F value  Pr(>F)  
## H_S               2 414.00 207.000  3.7619 0.08733 .
## season            2 356.36 178.182  3.2382 0.11122  
## Year              1  15.99  15.988  0.2906 0.60926  
## Monitoring_Route  1  61.50  61.499  1.1177 0.33112  
## Residuals         6 330.15  55.025                  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Still no variable’s p-value low than 5% in GBHE. But as the result of the fit and the alluvial diagram suggest, the independence variables have interaction.

# Two Way Factorial Design 
fit <- aov(value ~  season * H_S, aviaAllu[variable == "GBHE",])
anova(fit)

## Analysis of Variance Table
## 
## Response: value
##            Df Sum Sq Mean Sq F value  Pr(>F)  
## season      2 265.94 132.971  3.2609 0.11002  
## H_S         2 504.42 252.210  6.1850 0.03484 *
## season:H_S  2 162.97  81.485  1.9983 0.21622  
## Residuals   6 244.67  40.778                  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

With the involving of the interaction, the H_S variable seems to be significant this time. It means different season has different observation ways number.

2.4 Comparing with eBird

In this part, we want to compare the OWC bird data against eBird data. We want to keep the same five indicator spieces as our sample, which are CAGO, MAWR, GNHE, WOTH and WODU. This part mainly focuses on the different observation number of bird species between OWC data and eBird data which we find in the eBird website based on the week. With the limited source and access to eBird data, we only find some plot from eBird website which are related with these five indicator species. Then, we will make some plot of the same species based on OWC bird data.

In order to develop side-by-side comparison of the same details, we use week as our x-axis and observation number as our y-axis, which are the same meaning as eBird data. Then, we can analysis the total number of speices based on the week and see the comparsion betwwen OWC data and eBird data.

2.4.1 Canada Goose(CAGO)

2.4.2 Marsh Wren(MAWR)

2.4.3 Great blue hero(GBHE)

2.4.4 Wood Thrush (WOTH)

2.4.5 Wood Duck (WODU)

Based on the above plots, we can see that there are more data aboout total number of species in eBird dataset. However, with AvianMonitoring OWC dataset, there are lack of the observation number of species. Even we cumulate the three years (2016,2017 and 2018), it still cannot be covered in each week because of lack of data. Thus, we cannot make a very clearly comparison between OWC and eBird data.

3 Bald Eagle Nesting

Setting the working directory and reading the cleaned data:

setwd("C:/Users/indra/Desktop/DTD_Final/OldWomanCreek/Deliverables/FinalDemo/") # setting the working dir
eagle_raw_data <- read.csv("BaldEagle.csv") # reading the csv data

‘eagle_raw_data’ stores the whole dataset from csv file. We make data manupulations of the variable ‘eagle_raw_data’.

3.0.1 Barplot for Frequency

NestStatus = eagle_raw_data$NestStatus # select NestStatus from raw data
NestStatus.freq = table(NestStatus)
colors = c("red", "yellow", "green", "violet", "orange", "blue", "pink", "cyan")
barplot(NestStatus.freq, col=colors)

3.0.2 Frequency Distribution of Qualitative Data:

theme_set(theme_bw())
neststatus_subset <- subset(eagle_raw_data, select = c("NestStatus", "Temp", "Wind.Velocity.mph.", "Date", "CloudCover"))
g <- ggplot(neststatus_subset, aes(x=Date,y=Temp)) 
g + geom_point(aes(col=NestStatus)) + geom_smooth(method="lm", se=F) + labs(subtitle="Time Vs Tempearature", 
       y="Temparature (F)", 
       x="Date", 
       title="Scatterplot", 
       caption = "Source: BaldEagle")

Similarly, we want to

g <- ggplot(neststatus_subset, aes(x=Temp,y=Wind.Velocity.mph.)) 
g + geom_point(aes(col=NestStatus)) + geom_smooth(method="lm", se=F) + labs(subtitle="Temperature(F) Vs Wind Velocity(MPH)", 
       y="Wind Veloctiy (mph)", 
       x="Temparature (F)", 
       title="Scatterplot", 
       caption = "Source: BaldEagle")

eagle_raw_data$temp_z <- round((eagle_raw_data$Temp - mean(eagle_raw_data$Temp))/sd(eagle_raw_data$Temp), 2)
# above and below avg
eagle_raw_data$temp_type <- ifelse(eagle_raw_data$temp_z < 0, "below", "above") 
eagle_raw_data < eagle_raw_data[order(eagle_raw_data$temp_z),] #sorting

ggplot(eagle_raw_data, aes(x=NestStatus,y=temp_z, label=temp_z)) + geom_bar(stat='identity', aes(fill=temp_type), width = .5) + scale_fill_manual(name="Temperature", labels = c("Above Average", "Below Average"), values = c("above"="#00ba38", "below"="#f8766d")) + labs(subtitle="Normalised Temperature from 'BaldEagle'", title= "Diverging Bars") + coord_flip()