When working with time series data, it is good to know the data characteristics before conducting any analysis. Knowing the characteristics is useful in the following several ways:
- Data quality: Identify errors or inconsistencies in the data. If the data type is supposed to be numeric but there are non-numeric values, it can indicate a data quality issue.
- Understand data: Know the interval and frequency of the time series to understand the underlying process generating the data. If the time series is hourly, daily, or weekly, it can indicate the level of granularity of the data and the potential patterns or trends.
- Analysis methods: Determine the analysis methods suitable for different data types. If data has a seasonal pattern, a seasonal decomposition analysis can be appropriate. If data has a trend, a regression analysis can be useful.
- Outliers: Know the minimum and maximum values of the data to identify outliers or unusual data points that may require further investigation.
Visualization: Design appropriate data visualization techniques to effectively communicate the insights to stakeholders.
TD_SINFO returns one row for each series instance found in the target table. Each returned row provides the following information about the series:
- Index data type
- Starting index value
- Ending index value
- Number of series entries
- Indicator that the series is regular (discrete) or irregular
- Sample interval for regular series or average sample interval for irregular series
- Content type
- Minimum sample magnitude
- Maximum sample magnitude
- Average of magnitudes in the series
- Root-mean-square for each magnitude