Back to Research Hub

How-to guide

RORY - Replace the Demo Data with Real Shell Data (Step-by-Step + Prompt)

RWRaylen Williams · Oct 6, 2026
rorydashboardreal-datastepsprompt
Summary

Your current FluxView build works, but it shows sample numbers. This guide walks you through swapping in the real Shell data, what to fix, the prompt to use, and exactly where to upload your result in ProjectPulse.

Key findings
  • STEP 1 - Get the real data: Shell emailed the 8 daily CSV files to the whole team. Use the files whose names end in "daily_2023-05-10_to_2026-09-15.csv" (towers 001, 002, 003, 004, 005, 006, 013, 014). Keep them on your computer. Do not upload them to GitHub.
  • STEP 2 - Get the code into the shared repo: accept the GitHub invite from Raylen, clone the repo, create a branch called rory-real-data, and copy your current FluxView code into it.
  • STEP 3 - Read the guide first: open the Flux Tower Data Dictionary and the Quanterra Data Guide (both in Resources and in Research). Check the QC flag values before you filter anything.
  • STEP 4 - Load the real data: use the prompt below in your AI coding tool. It tells the tool exactly what to build. Review what it writes before you accept it.
  • STEP 5 - Fix the tower labels: towers 002, 003, 006, 013, and 014 get their names from the prompt. Towers 001, 004, and 005 show "Plot label pending" until Shell confirms them. Do not guess those names.
  • STEP 6 - Fix the air temperature unit: the CSV is in degrees Celsius, so show °C (you can add °F in brackets). Do not show °F alone.
  • STEP 7 - Remove the fake numbers: take out the "97%", "2,794 valid" and "Within target range" text, and any "since 6 PM" or "stable over 3 hours" text, unless the number is calculated from the real data.
  • STEP 8 - Test it: check every tower page, check the 30-day trend uses real daily values, and check the page at phone width (no sideways scrolling).
  • STEP 9 - Upload your result in ProjectPulse: open the Resource Center, click Add resource, enter the title "FluxView - real data build (Rory)", paste your live site or repo link in the Link box, type 2 to 3 sentences describing what changed in the Description box, choose "Other" for the type, and click Add resource.
  • STEP 10 - Update your task: open your task "Rory - replace demo data with real Shell CSVs" (or "Design the readings table..."), paste the same link at the bottom of the description, drag the card to Review, and post in the group chat that it is in Review.
  • PROMPT (paste into your AI coding tool): "This dashboard currently shows demo data. Replace it with real data from 8 flux tower daily CSV files, named US_TX_<tower>_daily_2023-05-10_to_2026-09-15.csv for towers 001, 002, 003, 004, 005, 006, 013, and 014. Each file has a column-name row followed by a units row. Skip the units row when parsing. Treat the text NA as missing. Read the QC flag column that goes with each variable, and count valid, suspect, and missing rows per tower; show those counts in the data-quality panel instead of hard-coded numbers. Show AIR_TEMPERATURE in degrees Celsius with Fahrenheit in brackets. SOIL_MOISTURE is a fraction, so show it as a percent. Label towers as: 002 Pasture-Treatment-1, 003 Sorghum-Treatment-1, 006 Sorghum-Treatment-2, 013 Sorghum Control, 014 Pasture Control. For 001, 004, and 005, show Plot label pending. Do not invent names. Remove hard-coded demo values and any text that says within target range, stable, or since a time, unless it is calculated from the data. Keep the layout mobile-friendly. Before changing anything, list the files you will change and why, then explain each change in plain language."
  • WHAT TO DO IF STUCK: post in the group chat with a screenshot and the exact error text. Try at most twice on your own first.

Part of the Shell NET–PVAMU Research Hub.