# 2D heatmap The whole size distribution over time, as a colour map: time across, particle diameter up, concentration as colour. The quickest way to see *which sizes* an event produced, rather than just how much. ```{figure} ../_static/gui/tab-heatmap.png :alt: The 2D heatmap tab with a total-concentration panel above a time-versus-size colour map :width: 100% Total concentration on top, the time–size distribution below. Reading up a column tells you the size distribution at that moment. ``` **Available for:** size-resolved (2D) datasets only — the tab is absent for single-channel instruments. ## The two panels The **top panel** is total concentration over time — the same trace as the [Time series](time-series.md) tab, for orientation. The **bottom panel** is the heatmap. Each column is one timestamp's size distribution. A rise that appears only in the lowest rows is a fresh, fine-particle source; one spread across the upper rows points at coarser dust. Both panels share a time axis, so zooming one zooms the other. ## Show as The distribution basis for **display only** — number (`dN`), mass (`dM`), surface area (`dS`) or volume (`dV`). The dataset itself stays as number. This changes the picture more than you might expect: a number distribution is dominated by the smallest particles, a mass distribution by the largest. Mass, surface and volume all depend on the particle density set on the [Metadata](metadata.md) tab. ## Normalize (dx/dlogDp) Divides each bin by its width in log-diameter space. Size bins are not equally wide, so without this a wide bin looks more concentrated simply because it collects more particles. Leave it **on** to compare bins fairly; turn it off to read the raw per-bin concentration. ## Colour scale **Log color scale** spreads several decades of concentration across the colours at once, which is normally what you want — aerosol concentrations vary by orders of magnitude, and a linear scale tends to show one bright band and nothing else. It needs positive values, so zeros are floored. **Color min** and **Color max** pin the scale to fixed limits; leave blank for automatic. Fixing them is how you make two datasets comparable — otherwise each one auto-scales to itself and the colours mean different things. ## Other controls **Log Y (conc.)** log-scales the y-axis of the top total-concentration panel. **Show activities** shades the marked periods on the top panel. ## Under the hood This is the two-panel figure `plot_timeseries` produces for size-resolved data: ```python import aerosoltools as at data = at.load_ops_file("measurement.csv") data.plot_timeseries(log=True, mark_activities=True, dtype="dN") ``` See [5 — Plotting](../examples/05-plotting.ipynb).