/appsilon-biodiversity

Appsilon coding assignment

Primary LanguageR

Species Tracker

This is a coding assignment for Appsilon, by Andrew McNeil.

Data Pre-processing

Data was split by country and saved as separate csv files. The original file was ~20GB, far too large to use in a shiny app. By splitting the data by country and removing unused column the data size can be greatly reduced.

sudo npm install -g @aleung/csvsplit
csvsplit -c 22 -o \
  ./preprocessed-data \
  ./biodiversity-data/occurence.csv

App Flow

  • A list of countries is determined by the file names
  • This populates the country selector
  • The species occurrence data is filtered by country
  • Based on this filtered data, the species selectors are populated
  • The scientific or vernacular name can be chosen
  • Whichever is chosen will automatically update the other to keep them in sync
  • The data is filtered again for a particular species
  • Based on this data, the date range selector is populated with the min and max observance dates in the data
  • The user can then optionally filter by date
  • The final data is displayed on the map and bar plot

Notes

  • Some values of vernacularName were missing. For this app I chose to remove these values, but if they were kept modification to the app logic would be required.
  • There were also duplicate rows in the data, which I removed. I don't have enough context about the data to know why this would be.
  • This app is responsive and should work on mobile. However, there seems to be a bug with plotlyOutput, and the plot is resizing incorrectly, throwing off the layout of the entire app. I would need to investigate further to solve this issue.
  • This app includes multiple uses of updateSelectInput. To fully test the modules which update the UI, shinytest should be used to test these portions of the app once the UI is stable.

Extras

  • Extra CSS has been used for the layout and custom bootstrap styling using the bslib package.
  • Performance has been optimized to a reasonable degree with the pre-processing of the data into partitions by country.