Time Series Explorer — a browser-based viewer that reads PyPSA netCDF results

19 views

Skip to first unread message

Franco Ferrucci

unread,

Jul 29, 2026, 9:30:50 PM (2 days ago) Jul 29

to pypsa

Hi all,

I use PyPSA for my research, and I kept wanting a fast way to just look at results — before writing a single line of plotting code. So I built one, and I'd really like your feedback.

Time Series Explorer → https://ferrucci-franco.github.io/timeseries-explorer/

Drop a PyPSA .nc file on it and it is parsed and organised the way you already think about your network: Buses, Generators, Lines, Links, Loads, Stores…, each with its time series (p, p_max_pu, p_min_pu, …) and its static attributes. No install, no notebook, no code.

From there you get analysis views built for energy data:

Integral — energy totals over any selected window (MW·h, auto-scaled to kW·h / GW·h / TW·h), as bars + share pie. Dispatch mix, curtailment and utilisation at a glance.
Profile — mean daily / weekly / monthly / yearly profile with ±1σ bands, all days or split by day type.
Heatmap — day × hour or week × day, so seasonality and outliers jump out immediately.
Plus Fourier, histograms, missing/NaN inspection, stacking, second Y axis, 2D/3D, and A|B comparison between two files.

And you are not limited to what is stored in the file: you can build derived variables on the fly and apply signal-processing transformations — derivative, integral, filtering/smoothing, resampling, gap filling for missing values, scaling and offsets, and arithmetic combining several signals. So things like "net injection at this bus", "ramp rate of this generator", or a smoothed/cleaned version of a noisy series are one step away, without touching the original file or exporting anything to Python first.

Two things that matter:

Your data never leaves your computer. Parsing and plotting happen entirely in the browser — nothing is uploaded, and the site sets no cookies.
There is a standalone build that runs fully offline for machines without internet.

It also reads CSV, Parquet, MATLAB .mat, generic netCDF, OpenModelica/Dymola results and pandas pickles (and audio files, for entirely unrelated reasons).

MIT licensed — source and issues: https://github.com/ferrucci-franco/timeseries-explorer

All comments are very welcome. The app has a built-in Feedback button that packages your session locally so you can attach it to a GitHub issue in one click. Bugs, PyPSA components I've missed, features you'd want — I'm listening.