_edited_edited.png)
Introduction :
Tableau helps data analysts to predict the trend of variables without involving any complex calculations. Before diving into prediction and trend analysis in tableau, it is important to know few key concepts in 'Time series'.
Time Series :
Time series is a concept which is used to get insights about the variables getting affected over a period of time. Time series requires a large number of data to make it reliable. There are three important components in 'Time Series' namely seasonality, trend and noise.
Seasonality :
This component holds data that gets affected in regular intervals. For example, the sales of jackets increases just before the winter starts.
Trend :
It is a pattern which shows the movement of data over a long period of time. Seasonality is the smaller portion of data and Trend is for a longer period of time. There are various types of trends broadly divided into 2 categories namely stationary and non stationary trend.
Up Trend - This kind of trend is non stationary. If there is a pattern which shows a consistent increase. it is called an uptrend. For example the overall trend of the sales of masks has been increasing from the year 2020.
image source - https://unsplash.com/
Stationary Trend - If there is pattern which does not undergo any consistent change, it is called a standard trend. The nature of this trend is deterministic with no major fluctuations. Stationary trend gives an absolute clarity in trend and seasonality without any kind of variation between the mean and variance. To give a real life example, the fit watch always gives an appreciation the more you exercise. You are clear with the ongoing trend as it is consistent and stationary.
Down Trend - This kind of trend is non stationary. If there is a pattern which shows a consistent decrease, it is called a down trend. For example, the purchase of calculators have been consistently going down.
image source - https://unsplash.com/
Noise :
Noise is any factor that can destruct the trend of data. It would cause a difficulty for the data analysts to analyze the data and derive at a conclusion because of the constant disruption which takes place. An unstructured data and improper filtering can lead to noise and this ultimately results in data corruption.
Sales forecast charts in Tableau :
Instead of doing some guessing games in the upcoming sales, Tableau provides a steady forecast facility to derive upon details and plan in advance. Sales forecast charts in tableau can be prepared with the help of the type of product the firm decides to manufacture or sell.
You need to forecast the sales for next upcoming years with 4 years of data sales given. The following steps can be helpful.
Note : To perform the following steps please download the sample super store data set from Kaggle.com
This particular sales forecast will be as per the region and segment. This is because you should have the products ready as per the segment and to which warehouse the products have to be sent.
Drag order date to columns.
Drag segment and sales to rows.
Drag region to colors.
You will arrive at a graph showing regional segment wise graph showing sales for 4 years.
To forecast the sales for next year
Go to 'Analytics' in the left side.
Under 'Model' you will have forecast.
Now you will be able to view the sales prediction for the upcoming business years.
Sales forecast can be done for next 2,3 years but it is not advisable to do so. It is not advisable because further forecasts are made upon the predicted chart we arrived at and not based on the actual results.
Steps to find the co relation between profit and sales respective of category and subcategory
What is 'Trend Analysis' ?
The study of co relation between the variables like sales, profit etc. is called trend analysis.
What are 'Trend Lines' ?
The co relation between two variables is figured out with the help of trend lines in Tableau. This helps us to observe the trend of the variables simultaneously. In this chart we are preparing we are yet to observe the trend between the profit and sales on the basis of category and sub category.
Drag profit to columns.
Drag sales to rows.
Drag category into color.
Drag subcategory into details.
Now we have derived at a chart called a scatter plot. A scatter plot chart shows the relationship between the values in x axis and y axis. The plot points in the chart show the relationship between the variables sales and profit in x axis and y axis. Note that each category belonging to one sub category will be of similar color.
Next, to show the co relation the concept of 'Trend line' is used.
Go to Analysis
Trend lines
Show trend lines
The co relation is displayed with the help of trend lines in the below given graph.
A 45 degree line in the graph shows a higher co relation.
A parallel line or a straight line represents zero co relation
Please note that a very much co related line also exists.
Note - The above given snapshots are from tableau and it can vary from one data analyst to other depending on their interpretation and preferred mode of visualization.