1. Course wrap-up
Welcome back! Let's close with a summary of what we learned throughout this process.
2. Time series data fundamentals
To start this course, we looked at the basics of what makes time series data. Time series data has a total span, the total length of data, and an interval, the amount of time between data points. That data can vary in a variety of different patterns, and can depend on a season of the year, a specific date range, or have an irregular pattern.
After that, we looked at how Power BI provides us capabilities to manipulate and work with time series data. DATEADD allows us to add or subtract a number of time periods, which helps with relative data analysis. DATEDIFF provides the amount of time that passes between two dates, and FORMAT allows us to change how a date is presented.
3. Window functions
Once we started working with time series data, we began to see the power and insights that we can gain. We began looking into the two major kinds of window functions.
Expanding window functions are those that have a single anchor point in time, and look at all all data before or after that point. These were macro-level analyses and typically revolve around an important date.
Rolling window functions analyze a specific amount of time, but will change which dates they look at as we get new data. These are great for analyzing the current state of our observations.
These analyses used functions like PARALLELPERIOD, SAMEPERIODLASTYEAR, and the period-to-date functions like TOTALYTD.
4. Forecasting time series
Moving past window functions, we began to explore predictive analytics through the use of time series forecasting.
Forecasts are predictions into future dates, and hindcasts attempt to "predict" our existing data. Both of these allow us to predict values within a confidence interval, which provides a range of values to expect in the future.
Power BI has built-in capabilities to quickly create a forecast, but we can also use DAX to do more programmatic approaches to forecasting through the use of averages and other statistics like geometric means. While Power BI isn't an industry standard tool for forecasting, it can be used for integrating these forecasts easily into existing reports.
5. Congratulations!
Congratulations on completing the Time Series Analysis in Power BI course! We hope that the content provided here will help you in your analytics journey, and we wish you all the best!