-
Notifications
You must be signed in to change notification settings - Fork 0
Frontend Pages
Static homepage with quick summary of the project and links to the tool’s main features.
Contains more detailed information about the project, its history, and contact information for how to get in touch with the team.
On this page, we show the historical trends of the three key variables of interest: number of pine beetles per two weeks, number of clerids per two weeks, and number of spots observed.
The chart will show these three series over the course of the selected years. The user may also choose to narrow in on a state, or on a forest within a state. If there are more than one forest selected, we take the average for that year across all of the available forests, given the filter selection. A box will appear underneath the chart which gives average and standard deviation for each variable across the entire period.
The map will show the average number of spots for a certain forest over the selected time horizon. The severity of the redness for each forest is determined by this average.
When the user changes a selection in the top menu bar, the chart and map adjust. The date range boxes have validation schemas built in -- you cannot select a year not available in the dataset. The system will also remember your last selection so that the next time the user returns, the same selection will be selected.
Finally, a user can export the data to a CSV. The CSV will only have the data which meets the current filter criteria.
The predictive model is a two-step process. First, the user must select a year and state of interest. By default, the current year is selected. Once the user selects a state, the system will calculate the predictive model results for each forest in the state and display this information in a table. On the map, the forest will be colored based on the probability of > 53 spots.
Please note that it takes a few seconds to run the predictive model for each state, so this process might take a while, depending on the number of forests in a state.
Second, the user must select an individual forest to drill down to see the full model results for that forest. By default, the system will automatically select the first forest in the list. The user can then click a forest on the map or in the table or select a forest from the dropdown above. They will then see the full output of the model, which includes the relevant inputs to the model, summary statistics about the output, and a chart outlining the various probabilities. Finally, the user can also conduct “what-if” analysis, by editing the inputs to the model.
There is also a download button in the left hand corner of the map on the "Predictive Model" page allows users to download the currently displayed map as a PDF (see image on next page for an example map PDF). When the button is clicked, the current state and year selected in the dropdowns above the map will be displayed in the downloaded map and used to name the file. Currently, the only data that can be downloaded is the probability of any spots occurring in the selected states. This is reflected in the downloaded PDF title.
There is also a hidden page located at https://pine-beetle-prediction.surge.sh/#UploadSurvey123Data. There is no link to this page and it should only be accessible by typing it into the browser. This page is password protected. Only admin users have access.
On this page, an admin user can select a state and forest from which to load data. The admin user should use this page to ingest the data for a certain forest only once the forest is done collecting for that year. This is because our ingestion process calculates input values to the predictive model for the forest by aggregating data from the forest’s traps. If a forest has multiple traps and not all traps are done collecting data, our system will run the calculations on incomplete data. Since these calculations are computationally complex, we did not implement the ability to update previously-existing calculations.
If subsequent edits need to be made, then the admin user should use the mLab website to directly edit the database.
Future improvements on this system should focus on loosening this restriction.
Once the system is done ingesting data from Survey123, it will show the data that was ingested with an option to download the data into a CSV for further analysis.