watermon-app

Florida Commercial Watermen's Conservation (FCWC) citizen science water quality monitoring application

Please visit https://marinebon.org/watermon-app for the output website dashboard.

process

• drop at least 3 casts (csv ea)
• only want downward casts, not return or noise in beginning

next steps

1. smooth the data per cast by interpolating into 1-foot bins. The data are pretty noisy and this would clean up the profiles.
2. Drop the dots in the profile plot. Unnecessary clutter.
3. Drop single cast transects from dropdown menu
4. Jitter the locations by 1/4-degree with thick line. Apply to all markers at once so shifted, but shape remains. To maintain confidentiality of fishing locations.
5. Figure out any other oddities with input data, eg multiple boat transects on single day for https://marinebon.org/watermon-app/date_2019-12-17.html.
6. Turn on Github Actions to regularly (eg daily), render website, which will involve a few more tasks:

• use docker image of Rmarkdown
• store Ben's Oauth Google Drive token to access files
• email Brendan & Ben (and potentially others: Chris Kelble) if it fails, or changes status

done

1. Update color ramps:

• consider colors used in this publication: https://tos.org/oceanography/article/true-colors-of-oceanography-guidelines-for-effective-and-accurate-colormap
• Temp: black-purple-gold
• Oxygen: decreasing white to black; <= 2 mg/L red scale below
• Salinity: blue-green-yellow
• Chlorophyll: white to green

1. Reduce silly decimal precision in contours of plots
2. 3 panes: 1) map, 2) 2 vars. 3) 2 more vars per 9-1-19_Venice.jpg
3. Add contour plots:

• Salinity
• Chlorophyll

1. reorganize data to go from wide to long:

   • data.csv: datetime | value | variable | units | sensor_sn | sensor_model
   • by default pull from sensor_model == "Aqua TROLL 600 Vented"

2. which temperature? want in water "Aqua TROLL 600 Vented" vs air temp from "In-Situ Bluetooth Device"

   • Instrument Properties Device Model = Aqua TROLL 600 Vented Device SN = 676038
   • Instrument Properties Device Model = In-Situ Bluetooth Device Device SN = 674484

3. oxygen_mgL = RDO Concentration (mg/L)

4. Go with Rmarkdown to Github Pages a la ship-cards.

• vs Enable shiny with flexdashboard

1. Add Dropdown to select date
2. Host on iea-demo.us, while exploring other long-term hosting options.
3. (Allow user login for upload.)
4. Operationalize processing of data:
   • setup cron to run every 5 min to check for new data in googledrive, fetch and process
   • Use googledrive with a service account token (see drive_auth() • googledrive)
   • When encountering any errors, send email with error log to user list using blastula for html-formatted emails; or even possibility of sending text messages via email through mobile phone provider: "Thanks for the data, but we're missing lon/lat. Can you help us tracking down these locations?"
   • Alternatively, the user could submit an email with attachment, then this cron job script processes the email attachments, adds to google drive, runs QA/QC and processes the file, a la robolimpet thermal sensor fetch of email (https://marinebon.org/thermal-data).
5. Read rest of data from Google Drive (see below).
6. Update color ramps (see 9-1-19_Venice.jpg):

• Temp: blue-red
• Oxygen: decreasing white to black; <= 2 mg/L red scale below
• Salinity:
• Chlorophyll: white to green

1. Plot all tracks and allow user to select bounding box and date slider? similar to Crosstalk
2. Smooth CTD cast along depth difference (ie dtime)
3. Update data live by reading Google Drive
4. Iterate based on user feedback.
5. theme: nik01010/dashboardthemes: BETA: custom theme support for R Shinydashboard applications.

Most common device to use (smartphone vs desktop)?

protocol

See VuSitu Operating Instruction.docx

• Aqua TROLL 600 Multiparameter Sonde - In-Situ

done

1. Avg lon/lat per csv (drop into yml), profile along multiple csv's.
2. Download US Coastal Relief Model - Floria and Eastern Gulf of Mexico for given area and just use that.

to explore

• Interpolation scheme?
  • PlotSvalbard::section_plot(interp_method = "mba")
  • "mba" for multilevel B-spline interpolation using the mba.surf function. Appears to produce the best looking results. Recommended.

data

Here is the google drive link: FCWC

BB local: /Volumes/GoogleDrive/My Drive/projects/iea-auto/regions/gm/fisher-monitoring/data/FCWC

The folder labeled 'raw data' is the one you want. This is the historical data.

Joe Bishop, the engineer working on this project and has setup the field units has another set of folders that the newly collected data go into. The following links are the newer data:

• Fort Myers
• Matlacha 2
• Naples
• BG (Matlacha)
• Venice
• Matlacha-2

data processing

• imaginaryfish/FCWC-data-processing

questions

1. Which SN for Temperature?
2. Where to maintain long-term? Want on NOAA site. Where to host Shiny server?

resources

• Red Tide Current Status | FWC
  • Red Tide Current Status

rmarkdown awesomeness

• What is R Markdown? on Vimeo
• https://rstudio.com/resources/cheatsheets/
• RStudio Webinars
  • The Essentials of Data Science