logo



Raspberries: Tiny, tart, and temptingly delicious. But before you pop them in your mouth, what chemicals have been used to get them to grow?


Project Summary

Data Sources

The data was collected from the U.S. Department of Agriculture National Agricultural Statistical Service.


  1. Data on the Chemical Agents used in California, Washington, and Oregon to grow strawberries and raspberries.

  2. Data on the Total Production per year or raspberries in California, Washington, and Oregon.


Data on Chemical Usage:

The original data on chamical agents contained a number of variables which included: pollination used, yield produced, fertilizer used, as well as restricted and non-restricted use of chemical fungacide, herbacide, and insecticide. The data was stored online and downloaded as a .csv file.

Chemical Data: https://quickstats.nass.usda.gov/results/6E4F3616-5CDA-34DF-B0D1-46D4BE03321E


Data on Raspberry production:

For total raspberry production by state, data was also collected from the U.S. Department of Agriculture National Agricultural Statistical Service. Please find links below:

For California: https://quickstats.nass.usda.gov/results/A82506DE-F3B2-3BDB-87D7-08B727A5C756

For Washington: https://quickstats.nass.usda.gov/#40159E97-BAE5-347A-B464-10DBC8AACBC5

For Oregon: https://quickstats.nass.usda.gov/results/8557227A-65F5-376E-8681-9A2D832003D8


For Symbol and Select Data Descriptions:

https://www.nass.usda.gov/Data_and_Statistics/Pre-Defined_Queries/ChemUseSymbolandDataItemDefinitions-fruit-vegetables.pdf




The Data: Columns and Values


##  [1] "Ag District"      "Ag District Code" "Commodity"        "County"          
##  [5] "County ANSI"      "CV (%)"           "Data Item"        "Domain"          
##  [9] "Domain Category"  "Geo Level"        "Period"           "Program"         
## [13] "Region"           "State"            "State ANSI"       "Value"           
## [17] "Watershed"        "watershed_code"   "Week Ending"      "Year"            
## [21] "Zip Code"
## [1] 211
## [1] 1990 2019


Explanation of Chemical Usage Dataframe


## [1] "RASPBERRIES"  "STRAWBERRIES"
##  [1] "CALIFORNIA"     "FLORIDA"        "WASHINGTON"     "OREGON"        
##  [5] "MICHIGAN"       "NEW JERSEY"     "NEW YORK"       "NORTH CAROLINA"
##  [9] "PENNSYLVANIA"   "WISCONSIN"
## [1] "RASPBERRIES, BEARING - APPLICATIONS, MEASURED IN LB" 
## [2] "STRAWBERRIES, BEARING - APPLICATIONS, MEASURED IN LB"
## [1] "CHEMICAL, FUNGICIDE"   "CHEMICAL, HERBICIDE"   "CHEMICAL, INSECTICIDE"
## [4] "FERTILIZER"
## [1] "CHEMICAL, FUNGICIDE: (AZOXYSTROBIN = 128810)"                         
## [2] "CHEMICAL, FUNGICIDE: (BACILLUS AMYLOLIQUEFACIENS STRAIN D747 = 16482)"
## [1] 1992 2019
## [1] 1990 2000



Data Cleaning and Scrubbing: Separate and Filter

Clean up was needed to filter out raspberries and details on chemicals used.


Data cleaning on the following collumns:

  • Commodity - to filter out Raspberries
  • Data Item
  • Domain Category - various chemicals


Clean up consisted of:

  • Separating values contained in the Domain Category column and created new columns of those values (Substance, Agen, Agent Name).

  • The Value column:

    • Filtering out (D), (NA), (Z) entries from the Value column.

    • Removing those values that were not numeric and would not help in the data analysis. Rows containing the following values were removed: (D), (NA), and (Z).


# Filter out Raspberries and separate/clean and arrange columns
rberry_data_bb <- rberry_data %>% filter(Commodity=="RASPBERRIES") 
rberry_data_bb %<>%separate(`Data Item`, c("Berry", "Kind", "Units"), ",")
rberry_data_bb %<>% separate(`Domain Category`, c("Substance", "Agent", "Agent Name"))

# Removing values of (NA), (D), and (Z) in the Values column
rberry_data_bb <- rberry_data_bb%>%filter(Value != "(D)")
rberry_data_bb %<>% filter(Value != "(NA)")
rberry_data_bb %<>% filter(Value != "(Z)")

# The Value column entries were characters and had commas. Removed commas and changed characters to numeric:
rberry_data_bb$Value = as.numeric(gsub("\\," , "", rberry_data_bb$Value))



Final Dataframe on Chemical Usage:


Year State Commodity Units Substance Agent Agent Name Value
2019 CALIFORNIA RASPBERRIES MEASURED IN LB CHEMICAL FUNGICIDE BOSCALID 300
2019 CALIFORNIA RASPBERRIES MEASURED IN LB CHEMICAL FUNGICIDE CYPRODINIL 900
2019 CALIFORNIA RASPBERRIES MEASURED IN LB CHEMICAL FUNGICIDE FLUDIOXONIL 600
2019 CALIFORNIA RASPBERRIES MEASURED IN LB CHEMICAL FUNGICIDE MYCLOBUTANIL 300
2019 CALIFORNIA RASPBERRIES MEASURED IN LB CHEMICAL FUNGICIDE PYRACLOSTROBIN 200
2019 CALIFORNIA RASPBERRIES MEASURED IN LB CHEMICAL FUNGICIDE TOTAL 24000



Other Dataframes: Production of Raspberries in California, Washington, and Oregon

  • For total production of raspberries by year, dataframes for production from California, Washington, and Oregon were used.

  • The 3 dataframes were merged vertically with rbind() since all had same column names.

  • The dataframe shows the year and production total for each state.

  • Value indicates the total production in pounds that year.

  • Downloaded data and selected columns for final Production dataframe.


# Read in CSV files for Production Totals by State Dataframes
cali_prod <- read_csv("Cali_Rasp_Production.csv")
range(cali_prod$Year)
## [1] 2007 2019
wash_prod <- read_csv("Wash_Rasp_Production.csv")
range(wash_prod$Year)
## [1] 2012 2019
oreg_prod <- read_csv("Oreg_Rasp_Production.csv")
#range(oreg_prod$Year)


California: Years 2007-2019

Year State Commodity Data Item Value
2019 CALIFORNIA RASPBERRIES RASPBERRIES - PRODUCTION, MEASURED IN LB 114500000
2018 CALIFORNIA RASPBERRIES RASPBERRIES - PRODUCTION, MEASURED IN LB 135000000
2017 CALIFORNIA RASPBERRIES RASPBERRIES - PRODUCTION, MEASURED IN LB 143280000
2016 CALIFORNIA RASPBERRIES RASPBERRIES - PRODUCTION, MEASURED IN LB 169670000
2015 CALIFORNIA RASPBERRIES RASPBERRIES - PRODUCTION, MEASURED IN LB 225300000
2014 CALIFORNIA RASPBERRIES RASPBERRIES - PRODUCTION, MEASURED IN LB 258900000


Washington: Years 2012-2019

Year State Commodity Data Item Value
2019 WASHINGTON RASPBERRIES RASPBERRIES - PRODUCTION, MEASURED IN LB 81000000
2018 WASHINGTON RASPBERRIES RASPBERRIES - PRODUCTION, MEASURED IN LB 86200000
2017 WASHINGTON RASPBERRIES RASPBERRIES - PRODUCTION, MEASURED IN LB 78050000
2016 WASHINGTON RASPBERRIES RASPBERRIES - PRODUCTION, MEASURED IN LB 80110000
2015 WASHINGTON RASPBERRIES RASPBERRIES - PRODUCTION, MEASURED IN LB 74100000
2014 WASHINGTON RASPBERRIES RASPBERRIES - PRODUCTION, MEASURED IN LB 69990000


Oregon: 2012-2017

Year State Commodity Data Item Value
2017 OREGON RASPBERRIES RASPBERRIES - PRODUCTION, MEASURED IN LB 5400000
2016 OREGON RASPBERRIES RASPBERRIES - PRODUCTION, MEASURED IN LB 8550000
2015 OREGON RASPBERRIES RASPBERRIES - PRODUCTION, MEASURED IN LB 9390000
2014 OREGON RASPBERRIES RASPBERRIES - PRODUCTION, MEASURED IN LB 8650000
2013 OREGON RASPBERRIES RASPBERRIES - PRODUCTION, MEASURED IN LB 8000000
2012 OREGON RASPBERRIES RASPBERRIES - PRODUCTION, MEASURED IN LB 8750000


Creating of Additional Data Frames: Join, Group_by, Filter Operations

#Production Dataframes
# join the three separate dataframes on production to make one large production dataframe
prod_df <- rbind(cali_prod, wash_prod)  # 2007-2019
prod_df_new <- rbind(prod_df, oreg_prod)  # 2012-2019
prod_df_grouped <- prod_df_new%>%group_by(Year)%>%summarise(Total_Pounds = sum(Value))
prod_df_reduced <- prod_df_new%>%filter(Year>2011)

# Additional Rberry data frames
rberry_data_reduced <- rberry_data_final%>%filter(Year>2011)
oreg <-  rberry_data_final%>%filter(State == "OREGON")
rberry_data_grouped <- rberry_data_reduced%>%group_by(Year, State, Agent)%>%summarise(Total_Pounds = sum(Value))
rberry_data_grouped_big <- rberry_data_final%>%group_by(Year, State, Agent)%>%summarise(Total_Pounds = sum(Value))
rberry_data_grouped2 <- rberry_data_grouped%>%group_by(Year, Agent)%>%summarise(Total_Pounds = sum(Total_Pounds))

unique(prod_df_new$Year)
##  [1] 2019 2018 2017 2016 2015 2014 2013 2012 2011 2010 2009 2008 2007



Plots

Raspberry Production from 2007 - 2011 and 2012 - 2019:


  • Production data for all three states starts in 2012.

  • For the years 2012-2019, California produced the most pounds of raspberries.

  • California has reduced raspberry production over the years and Washington has increased its production of the commodity.

  • There is no data on Oregon raspberry production for 2018 and 2019.

  • Since there is not data on production for Oregon and Washington before 2012, we cannot say that California has always produced the most pounds of raspberries.







Chemical Agents

  • Comparing the total number of records showing chemical agents in the dataset from 1991 - 2011 with the years 2012 - 2019.

  • The amount of herbicide used before 2011 is much greater than after.

  • The plots show that after 2011, there are slightly more instances of insecticide recorded and fewer instances of herbicide. Although we are comparing 20 years of records to 7 years of records, this difference is notable.






2012-2019: Total Pounds for Chemical Agents Used




Analysis at State Level

Create new dataframes: Group data by Year, State, and Agent (from 2011-2019)

# Create state level dta
wash_agents <- rberry_data_final %>%filter(State == "WASHINGTON")%>% group_by(Year, Agent, State)%>%
  summarise(Total_Pounds = sum(Value))
cali_agents <- rberry_data_final %>%filter(State == "CALIFORNIA")%>% group_by(Year, Agent, State)%>%
  summarise(Total_Pounds = sum(Value))
oreg_agents <- rberry_data_final %>%filter(State == "OREGON")%>% group_by(Year, Agent, State)%>%
  summarise(Total_Pounds = sum(Value))



Pounds of Each Chemical Agent Used on a Yearly Basis


  • Washington has used fewer chemicals in recent years than previously and only used sulfer once in 2009.

  • Oregon has also used fewer chemicals in recent years than in previous years.

  • Nitrogen, Potassium, and Phosphate seem to be the chemicals most used for all 3 states.







  • For California, there is only chemical data for the year 2019:




Comparing California and Washington in 2019

  • There is no data for Oregon in the year 2019.




Comparing Oregon and Washington:






Some Additional Take-aways from the EDA


  1. The pounds of insecticide used over the years has remained consistent for Washington and Oregon.

  2. Washington has produced fewer raspberries overall but used more kinds of chemicals than California and Oregon. This may be because it needs more kinds of chemicals due to its environment.

  3. Although California does not use as many kinds of chemicals, it does use more pounds of chemical than Washington and Oregon. This may be because it produces more pounds of the commodity.

  4. Sulfur was used sparingly in the late 2000s. Normally very little is needed so this may be why there is very little use recorded.

  5. This EDA analysis would have benefited from fewer entries of (NA), (D) and (Z) in the Value column resulting in more accurate plots.



Resources


The following resources were used to complete these slides and this project:

  1. USDA National Agricultural Statistics Service

  2. To learn how to remove the comma from the Value Column: Thank you to R for Excel Users. http://www.rforexcelusers.com/remove-currency-dollar-sign-r/

  3. RMarkdown Cheatsheet

  4. R for Data Science - for plot ideas.