Quasi-two-dimensional visualizations of microstructure in the thermocline are created by processing χpod signals in a non-standard way. As the moored instrument is pumped by surface waves, the fast thermistors at each end trace out vertically overlapping paths, from which we produce 2–3 m tall swaths. A swath capturing the unmistakable form of Kelvin–Helmholtz billows provides a proof of concept—albeit by relying on what is a rare event in our dataset. More commonly, swaths exhibit steppy temperature fields during weaker turbulence, an abundance of small overturns during stronger turbulence, or some combination thereof. To examine this continuum statistically, we take a 13-month dataset from an equatorial χpod at 120 m deep and divide it into 53 000 swaths, each 10 minutes long. Swaths are ordered based on their associated buoyancy Reynolds number (Re_b) inferred from our standard χpod processing. Clear visual differences in swath characteristics arise when comparing across an order of magnitude or more (e.g., Re_b < 10 vs Re_b = 10–100 vs Re_b > 100). With the help of a convolutional neural network to semi-objectively pick representative swaths, we present our best estimates of what microstructure looks like in two dimensions as a function of Re_b.