hi everyone! i'm playing around with elementary's anomaly detection before i put them in my production pipeline. i have put together a small test case and i am struggling to understand the results. first, i created a table and filled it with test data in snowflake with the following SQL commands:

CREATE OR REPLACE TABLE anomaly_detection_test_data
(
    test_val int,
    loaded_at timestamp
);

INSERT INTO anomaly_detection_test_data
select 
    uniform(1, 5, random(10002)) AS test_val,
    dateadd(day, '-' || seq4(), current_timestamp()) AS loaded_at
from table(generator(rowcount=>10));

and i used the following configuration for me anomaly detection test:

version: 2

models:
  - name: anomaly_detection_model
    config:
      elementary:
        timestamp_column: "loaded_at"
    columns:
      - name: test_val
        tests:
          - elementary.column_anomalies:
              column_anomalies:
                - average
              time_bucket:
                period: day
                count: 1

i built the model which ran the test and the test ran fine, which is what i expected because i have yet to introduce any anomalies to the dataset. when i took a look at the elementary dashboard, i was very confused by the graph. the minimum and maximum datetimes within my dataset are 2024-04-30 174156.523 and 2024-05-09 174156.523, respectively. the anomaly detection chart in the dashboard shows my first time bucket as being 2024-04-17 to 2024-04-18. how is this possible when those dates do not exist in my dataset?