Feature Specifications

Main Features are split into the front and back-ends.

Front-End

Completed:

• Built front-end UI for visualizing historical data and running the predictive model
• Embedded Mapbox maps to view data geographically/spatially
• Added plots from Chart.js to visualize data numerically
• Developed upload data page (password protected for admin users)

Ongoing:

Historical Data Page

• The year selection on the historical data page can be converted into a slider for easier use
• The map could be improved to show more full-fledged data. Currently, it shows average spots over the selected period. There may be more informative metrics and map-based visualizations.
• The summary statistics could be more customized to the scope of the filters. The best nation-wide summary statistics may be different from local summary statistics.
• Include past predictions on historical data chart to compare with actual spots
• Use county and forest map shapes instead of dots

Predictive Model Page

• Refine the micro-interactions
• Hovering over regions on the map can trigger outlining or bolding of that region
• Improve fidelity and consistency of county/national forest shapes on the map

New Pages

• Live data-watching page: Some users expressed interest in being able to watch the progress of data collection
• CSV upload page: Just like the UploadSurvey123Data page, we can have another hidden password-protected page which could upload CSV directly into the mongodb database, without needing to run a command line script
• Predictive model management page: Another hidden password-protected page which can change the predictive model. Update the R script being run. Change the inputs and outputs of the R script that is expected by the application

Back-End

Completed:

• Create database to store trapping/spot data
• Upload historical data to database
• Develop pipeline to fetch new data from Survey123 and add to database
• Develop pipeline to fetch all data in database and send to front-end
• Create method to run R-script (predictive model) and feed to front-end

Ongoing:

• Data validation - error check new data from Survey123 and predict/locate outliers
• Survey123 Ingestion Process - the Survey123 ingestion process could be more lenient in terms of updating data. It would be a complex but achievable undertaking to build a system which could efficiently check whether values need to be update and update them. Achieving would enable a host of neat features such as automatic pulling of data.