Text Mining Tool

Instructions

This interface was all developed in R using mainly the shiny and rtweet packages. The tool is online on shinyapps’ repository at the adress https://adsoncostanzi.shinyapps.io/TextMining/.

Development

Open the project using RStudio.

Deployment

After running the docker container or executing app.R locally, access the application on localhost:3838.

Docker

Build and run the application as following:

$ docker build -t text_mining .
$ docker run \
 -e API_KEY="AAA" \
 -e API_SECRET="BBB" \
 -e ACCESS_TOKEN="CCC" \
 -e ACCESS_SECRET="XXX" \
 -e APP_NAME="YYY" \
 -p 3838:3838 \
 text_mining

Simplified docker deploy using docker-compose:

$ API_KEY="AAA" API_SECRET="BBB" ACCESS_TOKEN="CCC" ACCESS_SECRET="XXX" APP_NAME="YYY" docker-compose up

Note: in both cases, set environment variables with credentials information. Otherwise the container may exit.

Local

Install system dependencies:

• r-base
• libssl-dev
• libcurl4-openssl-dev

Install R packages:

• curl
• openssl
• httr
• rtweet
• lubridate
• emayili
• dplyr
• shiny
• twitteR
• qdapRegex
• tidyr
• tidytext
• shinydashboard
• shinycssloaders
• shinydashboardPlus
• shinyWidgets
• shinyhelper
• shinyjs
• stopwords
• wordcloud
• plotly
• echarts4r
• networkD3
• widyr

Create credential file as described below.

Inside TextMining/app folder, run Rscript app.R.

Credential File Example

Once you have the Twitter API autorization you must create an R file called twitter_credentials.R inside the app/credentials folder. The file must have 5 objects: api_key, api_secret, access_token, access_secret, and app_name.

The twitter_credentials.R file should follow the template below:

# set Twitter API parameters

api_key <- "XXX"
api_secret <- "XXX"
access_token <- "XXX"
access_secret <- "XXX"
app_name <- "xxxxx"