▶

Try it in your browser

TRY ONLINE

Start the free version of LOGADEC right from this page, with nothing to download or unzip. It is the same file as the distributed LOGADEC 1.0.58.

Files you open in LOGADEC are read only inside your browser and are never sent to this site's server (the LOGADEC page itself has no analytics tag). Display settings are stored in this browser.

Recommended: Google Chrome or Microsoft Edge on a computer. Phones and tablets work too, but some operations are harder on a small screen.

A

See analysis results with the built-in demo (no file needed)

BUILT-IN DEMO

Steps

Click "Agree and start" to open LOGADEC. It opens with practice data for analysis ("Analysis tutorial", 6 series) already created.
In the [Analysis] tab on the right, open the "Outlier", "Change point", "Freq. anomaly" and "Corr. break" sections and click each button (such as [Detect outliers]). The settings are already set for this data.

If no data has been created, click [Demo] at the top, choose "Analysis tutorial" as the Type, and click [Generate].

Answers

This demo has planted anomalies. Check them against each result table and the markers and bands on the chart.

outliers at 24/72/101s, change points at 42/88s, frequency anomalies at 58-72s, correlation breaks at 58-75s.

[Full diagnosis] finds the same outliers, change points and frequency anomalies. It does not find the correlation break, because it picks the partner series automatically (the most strongly correlated one). Check it with the button in the "Corr. break" section (Comparison series is already set to "Correlation break").

B

Try it with sample test data

SAMPLE DATA

CSV files of fictitious test data (100 Hz, 120 s) with planted anomalies. Load one and run the analysis to see whether the planted positions are found.

Download a CSV below.
Drag the CSV onto the LOGADEC window (or choose it with CSV/Excel), then click [Import] in the import window.
In the [Series] tab on the right, click [Select all] to show the data on the chart (nothing is shown right after loading).
In the [Analysis] tab, open the section named for each sample and click its button.

**Outlier detection**Finds momentary spikes by how far they stray from the local median.[Analysis] tab → "Outlier" section → [Detect outliers]All 3 planted spikes (23.4, 61.8, 97.2 s) found at 23.4, 61.8 and 97.2 s, with no other detections.Download CSV (177 KB)

**Change-point detection**Finds the moments where the level of a signal shifts.[Analysis] tab → "Change point" section → [Detect change points]Flow steps at 38.0 s and 81.0 s found at 38.0 s and 81.0 s.Download CSV (165 KB)

**Frequency anomaly**Finds spans where a frequency component grows stronger.[Analysis] tab → "Freq. anomaly" section → [Detect frequency anomalies]A 37 Hz component added at 64–78 s is flagged as 60.0–82.0 s.Download CSV (159 KB)

**Correlation break**Finds spans where two normally linked channels stop moving together.[Analysis] tab → "Corr. break" section → [Detect correlation breaks]The 52–70 s span where winding temperature stops following current is flagged as 48.0–70.0 s (usual correlation 0.88).Download CSV (259 KB)

Try it with your own data

Your own CSV, Excel and NI TDMS files open the same way. Files are read only inside your browser and never sent anywhere. Check your organization's rules before using work data.

For company or offline computers

On computers without network access or managed company computers, use the download version. It is the same file.

Download free (1.5 MB) →