Contributors: Bailey Lei, George J. J. Wu

Project URL: https://github.com/UBC-MDS/DSCI_522_blei7_georgejjw_coco

Are darker chocolate simply better chocolate? This project explored the relationship between chocolate darkness (cocoa percentage) and chocolate expert ratings.

This project observed a weak and negative Pearson correlation between chocolate darkness and expert ratings. Using a Welch's t-test, this project also observed that the mean chocolate darkness difference between low and high quality chocolate bars was practically trivial. There seemed to be no clear relationship between chocolate darkness and expert ratings. Limitations and future directions of this project were discussed.

The cocoa percentage and expert ratings of over 1,700 chocolate bars were sourced from Kaggle under the CC0 1.0 license.

For a detailed description of the data set, please see Kaggle.

1. Install and configure Docker.

2. Install the docker image for this project by using the terminal command:

docker pull wuj4979/mds-coco-analysis

Please note that this docker image is about 698mb in size.

3. Clone/download this repository, and navigate to the root of this project.

4. Reset the project to its initial state by executing the terminal command:

docker run --rm -v PATH_TO_THIS_REPO:/home/DSCI_522_blei7_georgejjw_coco wuj4979/mds-coco-analysis make -C "home/DSCI_522_blei7_georgejjw_coco" clean

Replace PATH_TO_THIS_REPO with the absolute path to the root of this project on your computer.

For Windows users, please use CMD or PowerShell to execute the command, and use Windows style path such as //C/Users/blei/Downloads/DSCI_522_blei7_georgejjw_coco.

5. Run the data analysis with the following terminal command:

docker run --rm -v PATH_TO_THIS_REPO:/home/DSCI_522_blei7_georgejjw_coco wuj4979/mds-coco-analysis make -C "home/DSCI_522_blei7_georgejjw_coco" all

Replace PATH_TO_THIS_REPO with the absolute path to the root of this project on your computer.

For Windows users, please use CMD or PowerShell to execute the command, and use Windows style path such as //C/Users/blei/Downloads/DSCI_522_blei7_georgejjw_coco.

A final report will be compiled at doc/final_report.md and doc/final_report.html.

If you don't have Docker installed, you can also run this analysis by using GNU Make as follows.

1. Clone/download this repository, and navigate to the root of this project.

2. Reset the project to its initial state by executing the terminal command:

make clean

3. Run the data analysis by executing the terminal command:

make all

Please make sure to have all the dependencies for this project installed before executing this command.

The make all command will automatically generate all relevant data files and figures for this project.

A final report will be compiled at doc/final_report.md and doc/final_report.html.

If you don't have Docker or Make installed, you can also run this analysis by using a shell script.

1. Clone/download this repository, and navigate to the root of this project.

2. Run the analysis by executing the terminal command:

bash run_all.sh

Please make sure to have all the dependencies for this project installed before executing the shell script.

The shell script will execute the following commands.

# prepare data
Rscript src/prepare_data.R data/flavors_of_cacao.csv data/cleaned_coco.csv

# generate figures
Rscript src/visualize_darkness.R data/cleaned_coco.csv results/darkness_dist.jpg
Rscript src/visualize_rating.R data/cleaned_coco.csv results/rating_dist.jpg
Rscript src/visualize_scatter.R data/cleaned_coco.csv results/scatter_darkness_rating.jpg
Rscript src/visualize_mean.R data/cleaned_coco.csv results/mean_darkness_rating.jpg

# generate tables
Rscript src/describe_darkness.R data/cleaned_coco.csv results/darkness_stats.csv
Rscript src/describe_rating.R data/cleaned_coco.csv results/rating_stats.csv
Rscript src/compute_pearson.R data/cleaned_coco.csv results/cor_pearson.csv
Rscript src/compute_ttest.R data/cleaned_coco.csv results/ttest.csv

# compile final report
Rscript -e 'rmarkdown::render("src/generate_report.Rmd", output_file = "final_report.md", output_dir = "doc")'

Here is a brief description of each script.

Here is a flowchart depicting the relationships among the scripts.

Upon executing the above scripts, a final report for this analysis will be generated inside the doc folder under the filename final_report.md and final_report.html.

• GNU Make (≥ 4.1)

• R (≥ 3.0.2)

• R packages:

  • tidyverse (≥ 1.2.1)
  • readr (≥ 1.2.0)
  • infer (≥ 0.4.0)
  • knitr (≥ 1.17)
  • rmarkdown (≥ 1.7.4)
  • here (≥ 0.1)
  • testthat (≥ 2.0.1)