Uploading datasets¶

FinSentry takes one Excel export at a time and walks you through a short review before it processes the data.

What the file needs¶

• An .xlsx file of at most 50 MB. Other formats are not accepted.
• Data on the first sheet, with column headers in the first row, and at least one row of data.

Steps¶

1. Upload¶

On the Upload dataset page, drop the file in or click to browse. FinSentry stores the file and suggests a role for every column — cost center, time period, or metric — along with a unit and a readable name, based on the first rows.

2. Choose dimension and time columns¶

Confirm which column is the cost center and which columns hold the period. These decide how every metric is grouped and dated. Choosing Start import imports the rows in the background.

3. Rename metrics and attributes¶

Set the names people will see across the dashboard, pivot, and anomaly views, and link each % metric to its absolute value. You can edit while the import runs; Continue to formulas becomes available once it has finished.

4. Set formulas and pick anomaly metrics¶

For each metric, define how it is calculated (or leave it as the raw column), and tick the metrics you want scanned for anomalies. See Adding formulas.

Choosing Confirm and save starts anomaly detection.

After you confirm¶

FinSentry runs anomaly detection on the selected metrics and works out drivers for the metrics that have formulas. This usually takes under a minute. Progress is shown on the dataset page; results appear on the anomalies page, and the metrics appear on the dashboard, pivot, and cost-center pages.

Adding more months later¶

To extend an existing dataset with new months, use Append on the dataset page rather than starting a fresh upload. This keeps the history continuous, which matters because anomaly detection needs six earlier months to judge each metric. The appended file must have exactly the same columns as the dataset; if it does not, the append is rejected and the existing data stays as it was.