This is the sixth post in a series of blog posts using a theme of “Tools of the trade”. The series targets software tools, statistical concepts, data science techniques, or related items. In all cases, the topic will contribute to accomplishing data science tasks. The target audience for the posts is engineers, analysts, and managers who want to build their knowledge and skills in data science, particularly those in the Microsoft Dynamics ecosystem.
This post is the first in a series of posts exploring data science applications of time series.
What is it?
Time series data “is a series of data points indexed (or listed or graphed) in time order. Most commonly, a time series is a sequence taken at successive equally spaced points in time. Thus it is a sequence of discrete-time data.” (https://en.wikipedia.org/wiki/Time_series).
The simplest way to visualize time series data is a table with two columns. The first column is ‘Time’ and each row represents an equal increment of time from the previous row. The second column is ‘Value’ and represents what you are measuring for each time increment. Some simple time series include:
- Amount of rainfall each day
- Bike miles ridden each week
- Calories taken in each day
How do I use it?
Time series data shows up in many places at work and at home.
At work, data used to monitor and measure the use of our service is often in time series form. We measure active users on a daily, weekly, and monthly basis. We track system health at more granular units. We target key performance indicators on a weekly and monthly basis. Data measured at equally spaced intervals facilitates goal setting and measurement, forecasting, and anomaly detection.
The primary time series focus for me in the Dynamics 365 for Finance product in recent months are applications in budget planning and cashflow where time series forecasting can have significant value.
At home, the Garmin workout app that uploads data from my Garmin devices shows me a variety of information on a daily, weekly, and monthly basis – workout minutes, miles ridden or ran, hours of sleep per night, etc. Financial planning involves a budget tool that focuses on monthly increments of time.
Discussion
For the purpose of this and future blog posts on time series, I will be using my monthly electrical bill spending as the time series data to explore. Electricity is my family’s primary source of heat, hot water, and air conditioning. Living in a climate that has cold winters and warm summers provides lots of interesting variation in the data. I have almost four years of monthly time series data that I’ve graphed below.
The plot shows some strong seasonality in the monthly bills and what appears to be an increasing trend. It also shows that there is one month with missing data. That’s a common challenge with time series data and one we’ll tackle in the next post.
References
https://en.wikipedia.org/wiki/Time_series
Picture Details: Clouds on water, 6/2/2020, iPhone 7, F1.8, 1/1008 s, ISO-20