Parquet Viewer: Plot + Inspect Your CAN Bus Data Lake
This free online Parquet viewer lets you plot signals and inspect Parquet files from your CAN bus data lake - directly in your browser. No install needed, your files are never uploaded and you can also simply use it offline.
Try it on the sample data above, or drop your own data lake folder, zip or Parquet files on the viewer.
Need a quick way to view Parquet files from your CAN bus data lake?
This free Parquet file viewer lets you load an entire data lake (or single Parquet files) in your browser. You can plot your DBC decoded CAN/LIN signals over time - and inspect the contents of each file.
Key features include:
- Load a data lake folder, a zip of one or single Parquet files (also via drag + drop)
- Plot up to 10 signals from one or more messages of a device
- Use a shared Y axis, one Y axis per signal or stacked lanes (one per signal)
- Drag to zoom: The data is re-queried at a finer interval, like in Grafana
- Resample by mean, last value, min or max - and see CAN signals as step lines
- Hover readouts, a signal summary table and arrow-key panning
- Table tab with each file's columns, types, rows, row groups and compression
- Share the current view as a link and show times in local time or UTC
- Opens on sample data from a CANedge2 + CANmod.gps (EV drive)
The viewer is made for Parquet data lakes created from CANedge log files via our MF4 decoders - locally or in the cloud. Learn more in our Parquet data lake intro.
Many CANedge users review their Parquet files via Tad, a free desktop tool for tabular data. Tad is great for browsing rows and columns, but it shows the data as a raw table (like a CSV file) and opens one file at a time.
For CAN/LIN data, you want to see signals over time - across many files and days. This Parquet viewer plots signals from your entire data lake with Grafana-style zoom, while the table tab lets you inspect single files.
No install is needed either, which helps if IT restrictions block software installs on your PC.
Grafana dashboards and Python/MATLAB scripts are great for ongoing analysis, but overkill if you just want to take a quick look at your data. This is where the Parquet viewer fits in.
Yes. The viewer is a single, self-contained HTML file (about 13 MB) with no external dependencies. You can download it and open it in your browser from your local disk - without an internet connection or any installation. This makes it a simple Parquet viewer for Windows PCs with IT restrictions.
To use it offline, click 'download' at the top of the viewer and open the saved 'parquet-data-lake-viewer.html' file in your browser. We recommend Chrome or Edge.
Your data stays private, also when you use the online version. The viewer reads your files locally and queries them via DuckDB-WASM in your browser, so nothing is uploaded to a server.
Use the data menu in the top left of the viewer to open a folder, a zip or one or more Parquet files. You can also drag + drop them onto the viewer. Click 'Sample data' to return to the sample.
The viewer expects the CANedge data lake layout, i.e. device/message/yyyy/mm/dd/*.parquet. You can open the full data lake folder, a device folder or a single message folder.
The Parquet files must contain DBC decoded data. To create them, decode your CANedge MF4 log files via our MF4 decoders - or set up an automated Parquet data lake in Amazon, Google or Azure. For a cloud data lake, download the folders you need to your PC and open them in the viewer.
The viewer only reads the files needed for the current view, in batches with bounded memory. We have tested it on an 18 GB data lake with 100,000+ files.
Zips that unpack to more than 4 GB are refused (open the folder instead). Your browser may take up to a minute to list a folder with 100,000 files, and plotting months of a busy message can take seconds.
For fleet-scale analysis, we recommend a Parquet data lake with e.g. Grafana-Athena dashboards or Python.
Apache Parquet is an open, columnar file format. It is compact and fast to query, which makes it popular for storing large amounts of data for analysis.
A Parquet data lake is a structured collection of Parquet files in a local folder or the cloud (e.g. AWS S3). With the CANedge, our MF4 decoders DBC decode your log files into a data lake with folders per device and message.
The data lake can be queried via SQL tools like Amazon Athena or DuckDB, Python/MATLAB, or dashboards like Grafana. Learn more in our Parquet data lake intro.
We built the viewer with Claude Code and our CANedge AI skills. It is based on the open source libraries DuckDB-WASM and uPlot (see 'Licenses' in the viewer).
Help us create the best Parquet viewer: Please contact us with any feedback, suggestions or bug reports.
Need to log / decode CAN bus data?
Get your CANedge today!
