Autodesk Tandem Visualisation

 As I continue exploring how to display data for end-user visualisation (beyond facilities managers) — using Node Red UI templates and Python (Pandas + Matplotlib) — I’ve been reflecting on lessons from troubleshooting the Boukunde devices at the and connecting their outputs into Autodesk Tandem (see image below).

In June and July, I collaborated with researchers at the Council for Scientific and Industrial Research (CSIR) to understand why certain payloads were incomplete or missing from the API. At first, we needed to expand the data structure (adding Decibels, Lux, GPS, and device metadata), but the deeper issue emerged later: the devices themselves were not reporting to the interface server. This meant that while the Python scripts and API calls worked as expected (returning “0 readings”), no fresh data was flowing upstream.

Some key takeaways:

 1️⃣ Logs returning “0 readings” don’t always mean broken code — sometimes it’s a network or device-side issue.

 2️⃣ API payloads evolve, so pipelines need schemas that can flexibly handle new keys and null values.

 3️⃣ Real-time systems succeed only when hardware and software teams collaborate — digital twins depend on both ends working in sync.

This is what makes digital twin research so fascinating: aligning physical devices, networks, APIs, and cloud pipelines into coherent systems that support smarter building management. My next step is shifting focus toward how the display of environmental data through digital-twin-inspired visualisations can influence end-user behaviour.

(Using the term “digital twins” with caution — perhaps better to say drawing on digital twin methodologies).