A time series is basically nothing more than data points indexed over a certain time period. It’s an important data input to create reliable forecasting for your business. To be able to use time series for your business, you need a couple of years of data. The more data points, the better. The more specific, the more accurate.
The most basic time series data is a point in time with a value attached to it. For example: the number of units sold. The closing value of Nasdaq. The demand of one of your products.
Time series analyses will become more accurate if you add more data. What if you know your products’ sales have a high correlation with outside temperature. Even without having years or even decades of sales data, you could add decades of weather data to your data input. This will immediately make your input for the time series analyses more valuable and create a more accurate output to base your purchasing, manufacturing and logistics on.
How to do a time series analysis?
The easiest way to do a time series analyses is to just look at the graph. If you have your data in Excel, you can make a line graph based on the data. The most important question that has to be answered when analyzing the data is: Why!? Why is the number X on day Y and not Z on day Y.
If you only had number of sales plotted over time, you would have a hard time coming to any meaningful conclusions. But if you added more data points to the analyses it starts to make sense. When you add the temperature line to the graph it will change the conclusions you’d make. If you add another dataset, let’s say demand for a comparable competitive product, your predictions will really start to get accurate.
From analyses to time series forecasting
Forecasting is predicting. And everyone knows that the future is unknown. Unless you have enough data, no one will be able to predict the future. If you have enough models and datasets of the past, you can make pretty reliable forecasts of the future.
It’s often easier and more accurate to forecast for a shorter time horizon compared to a longer horizon. The further the point in time the less accurate forecasts usually get. If you want, you can frequently update your statistical model as you gain more new information that could help make more accurate forecasts.