General sensor and lab-data cleanup
Clean noisy engineering time-series data — without writing a pipeline
Inspect CSV or TSV metadata locally and request a non-binding workload estimate. In the account, select measurement columns and review access before processing. The one-off workspace returns the selected cleaned channels, a processing summary and an integrity manifest.
The estimate is free. File values stay in your browser during pricing; only file shape and selected outputs are sent.
Use the Lab CSV cleaner for one completed engineering file
This is the shortest self-service route for a CSV or TSV that needs cleaning without a recurring subscription or a custom pipeline.
- Best for
- Numeric engineering time-series with a time column and measured channels.
- Local first step
- Estimate the likely processing requirement in the browser before creating an account.
- You receive
- Selected cleaned channels, a processing summary and an integrity manifest. Locally saved files are available on the browser/device used for the run.
- Buying model
- One defined job, no subscription. The final scope and cost are shown before payment.
Free local estimate
Check the likely processing cost before creating an account
Choose one or several files. Your browser reads file size, row count and column count locally; signal values are not uploaded for the estimate.
Drag CSV or TSV files here or
Several files can be estimated together. Counting happens locally in the browser.
The estimate is not an order. Workload can be estimated while monetary pricing is unavailable; no price is invented in that state. The €15 CSV top-up minimum is not a file price. The workspace validates the upload and account access separately.
Input and result
A defined file-cleaning job
What goes in
Time-ordered numeric engineering data in CSV or TSV: test-bench logs, DAQ exports, climate-chamber runs, process measurements and sensor telemetry. Battery cycler files should use the Battery workflow, which adds battery-specific mapping, cycle semantics and evidence.
What comes back
The current one-off CSV workspace uses online A1 and returns the selected measurement columns with a time or row index, a processing summary and a source/output hash manifest. It is not a per-value change ledger. Use Filtration for other output modes and Battery for battery-specific reports.
Why automate it
Repeatable cleaning without a per-file script
Manual spreadsheets
Edits are slow to repeat and difficult to audit across files.
One-off Python
Flexible, but the pipeline, thresholds and report format must be rebuilt and maintained.
One fixed filter
A setting that suits a quiet region can smear transients or add error when the spectrum changes.
How it works
Estimate, run, download
Estimate locally
Select files and outputs. Only shape metadata is sent for the public estimate.
Confirm and launch
Sign in, review the validated price and add balance if needed. The job starts only after explicit confirmation.
Download the result
Receive the cleaned file and the accompanying processing and integrity records in your account.
Evidence
Hard signals are where a fixed filter can fail
RMSE reduction relative to the noisy input; higher is better and a negative value means the method added error. These are protocol-bound benchmark cases, not a guarantee for every file.
| Signal · SNR | Best listed classical baseline | Low-latency | Offline |
|---|---|---|---|
| Multi-tone vibration · 30 dB | −6.5 dB | +8.3 dB | +12.5 dB |
| Chirp · 20 dB | −11.7 dB | +5.3 dB | +0.0 dB |
FAQ
Before you process a file
Which files fit this workflow?
CSV or TSV with ordered numeric engineering measurements and a meaningful time or sample order. Business tables and unstructured documents are not the target.
Is the estimate free?
Yes. Local inspection and the public workload estimate do not create a processing job or payment. Monetary pricing is shown only when the service supplies it; access and usage are checked separately in the account.
What happens to unused payment balance?
The CSV workflow can use the lower €15 payment minimum. Only the job cost is consumed; unused top-up balance stays on the account and does not expire.
Are my values uploaded during the estimate?
No. Values are uploaded only after you sign in and explicitly launch a paid processing job. See the Privacy Policy.
Start
Estimate your files before you buy
Use the local estimator above, then open the account when you are ready to run the job.
Guides
Practical time-series cleaning
Remove noise from sensor data
Moving average, Savitzky–Golay, Butterworth, wavelet and Kalman approaches.
Clean battery cycler data
Why battery cleaning needs cycle and peak-aware handling.
Fill gaps in irregular time series
Five methods and where each one breaks.
Despike and remove outliers
Hampel and residual methods without deleting real events.