GetNuts

GetNuts is a pipeline that takes data tables (.xlsx) pulled from Pulsar and can output a series of static plots as well as an interactive shiny dashboard. This pipeline is tested to work on Windows, macOS and Linux.

To use GetNuts, download or clone the repo to a folder on your computer, then create a folder /Data where pulsar .xlsx data can be placed. Important: if using multiple excel files, it is important that only the first file has header information in the table.

GetNuts analysis can be modularized by selecting the desired outputs in config files, which enables separation of project and analysis. An example config file can be seen in the /config_files directory.

Installation

Dependencies:

• readxl
• quanteda
• magrittr
• tidyr
• ggplot2
• leaflet
• rgdal
• sp
• spdplyr
• ggmap
• rgeos
• maptools

Prior to running the Pipeline install:

install.packages(c(
  "readxl","stringi","quanteda","magrittr",
  "tidyr","ggplot2","ggpubr","shiny","leaflet",
  "sp","rgdal","spdplyr","rgeos","maptools"))

and

The latest ggmap must be installed from github:

if(!requireNamespace("devtools"))
install.packages("devtools")
devtools::install_github("dkahle/ggmap", ref = "tidyup")

Then, clone or download the GetNuts repository to your local computer.

Configuration

To analyze data, create a folder /Data in the repo parent folder and place Pulsar .xlsx data there.

To perform an analysis, the user must specify a config.txt file to run with GetNuts. Config files are stored in /config_files and specify parameters of the analysis to be performed. These include:

• input_file_dir: where the data has been stored (Default is Data)
• input_file_list: The files in the input_file_dir to be analyzed. Important: the first file should have the header information, all files after should not have header information
• sheet_name: sheet of the excel file to take the data from (Default is Sheet1)
• project_name: Name given to analysis. This will create a folder of this name in /output_files/ where all output from this analysis will be found.
• rerun_data_preprocessing: ("yes" or "no") Data Preprocessing, which involves stopword removal and stemming is the most computationally expensive process in the pipeline. If set to "yes", pipeline will be re-run from the beginning. (Default: "no")
• rerun_data_labelling: ("yes" or "no") Data labelling refers to when the preprocessed data is checked against the supplied dictionaries (in /Dictionaries). If a change is made to a dictionary, the analysis does not need to be re-run from the beginning, it can begin here if set to "yes". (Default: no)

For the outputs, the user can also specify in the config file what is desired and in which format.

• do_static_plots: ("yes" or "no") Specify whether to export static plot files (Default: yes)
• static_plot_format: Desired output format for output plots. Can be any one of png,jpeg,pdf,eps,svg,tif,bmp. (no default)
• launch_shiny_dash_board: ("yes" or no") To launch shiny dash board change this option to "yes".

Analysis

To run the analysis, open an R console and

• Navigate to the GetNuts Directory

setwd("your/GetNuts/Directory")

• Source the Main script:

source("GetNuts.R")

• Call the function allergies_and_social_media with the desired config file:

allergies_and_social_media("config_files/your_config_file")

Which will run the entire pipeline with the configuration supplied.

• Check output_files/$project_name/ for all outputs from the analysis.