Goal
- Confirm successful Terrier clone
- Integrate outside data for flight tracking and turbulence data
Workflow
This week has been my most successful thus far during my internship with Wet Dog Weather! I continued working on my aviation turbulence dashboard and its core components. Here are the main achievements I made this week:
- Confirmed a successful Terrier Clone
- Integrated OpenSky for nationwide flight tracking
- Integrated FlightAware for specific flight information and routing
- Integrated NOAA Graphical Turbulence Guidance Model Data from the surface to 40k feet
Once I successfully implemented those, I began working with the data. This included creating detailed info cards that displayed the flight’s route and other available information. I also asked Claude to create a 2D vertical profile of the selected flight, showing turbulence along its route.
While I wanted to keep implementing new features, this project is long-term, so I spent the rest of the week refining the features I mentioned above. The demo video and images below showcase the dashboard in its current state.
Moving forward, here’s what I plan on implementing:
- Terrier Radar Data
- Another turbulence model data source (potentially)
- Live METAR station plots at each airport
- PIREPS (Pilot reports on flying conditions)
As I continue to work on data integration and refinement, I’ll be brainstorming ideas for the passenger side of the dashboard. The main challenge will be figuring out the most effective way to communicate the data to the customer, assuming they know nothing about aviation or weather. This will be easier once I have a defined dataset.
Prompts Used
My prompts this week included many visual references. ForeFlight’s dashboard was a big inspiration piece for the turbulence data. I directed Claude Code to their website to determine where and how they obtained the turbulence data, which proved highly successful.
What Worked
- As mentioned above, Claude Code was able to “interrogate” ForeFlight’s website to learn how they gathered their data.
- Claude’s Fable model has been re-released for the public. What I noticed is that it produced far fewer errors, at the cost of longer thinking/working times and token usage.
What Didn’t Work
- Speaking of Fable, starting Tuesday, July 7th, it won’t be included in any subscription plans. It will be available only on a credit basis.
- During my sessions with Claude, it would repeatedly open its in-app browser to verify changes it made. This just wasted time and tokens, and I had to address it with Claude across multiple sessions.
Lessons Learned
A good lesson I learned this week was that data outside of Wet Dog Weather’s product line seems to merge well and complement Terrier data. This opens the door to plenty of opportunities and new, exciting ideas!
Images/Video
Image 1 (Click to enlarge)
Image 2 (Click to enlarge)
Image 3 (Click to enlarge)