Creating a Custom Series on a Chart (Trends)

Owned by Zoe Baldwin
Last updated: Jul 10, 2024


Overview

This feature is available in PhixFlow version 11.1.1+.

This page will show you how to add custom series to any PhixFlow chart such as quartile, trend, target and/or arbitrary lines. 

There are 4 types (modes) of custom series:

  • Algorithm
  • Percentile
  • Trend Line
  • Expression

For the Properties, see Custom Series.

Algorithm Lines

Algorithm lines involve applying mathematical formulas or algorithms to chart data. The lines can help identify trends, make predictions or highlight patterns in the data. The chosen algorithm will be calculated using values from each column of data.

Below is a line chart showing how the quantity of different objects changes year-on year from 2001 to 2014. The teal line represents the mean value for each horizontal axis attribute (year). The yellow line represents the cumulative percent. Click on the image to enlarge it.

Creating Algorithm Lines

  1. Create a line or bar chart, see Creating a Line Chart or Creating a Bar Chart
  2. Right-click the chart and click  Show the View Configuration
  3. In the  Properties, expand the Custom Series section
  4. Click the  icon to create a new Custom Series
    1. Name: Enter a name 
    2. Right Axis: Enable this to display the custom series for the data on the right axis
    3. Mode: Choose Algorithm
    4. Algorithm: Choose from:
      1. Cumulative Sum - shows the running total of the data values

      2. Cumulative Percent - shows the total percentage accumulated at each point in the data series

        Cumulative Percent reflects the cumulative contribution or impact of each percentage value as it progresses over time or through a particular set of data.

      3. Mean - shows the average value

      4. Minimum - shows the smallest value 
      5. Median - shows the middle value, when the data is ordered ascending
      6. Maximum - shows the largest value 

      7. Standard Deviation - shows how spread out the values are from the mean

        Standard Deviation helps to understand the dispersion or variability of the data points. A larger standard variation indicates that the values are more spread out from the mean.

      8. Sum - shows the result of adding all values together
        1. For a multi-series chart, this will be the sum for each column of data, for example: 


        2. For a single-series chart, this will be the sum of all values, for example:

    5. Colour: Choose a colour for the chart line from the colour picker
    6. Style: Select a line style from dashed, dotted or solid

Percentile Lines

Percentile lines are a type of quantile showing how a value compares to other values within the same set of data. They indicate the value below which a given percentage of values in a set of data falls, e.g. the 50th percentile is represented below the point at which 50% of the data falls.

Below is a line chart showing how the quantity of different objects change for each quarter of the year 2015 to 2016. The dashed yellow line represents the 30th percentile, while the dashed purple line represents the 60th percentile. Click on the image to enlarge it.

Creating Percentile Lines

  1. Create a line or bar chart, see Creating a Line Chart or Creating a Bar Chart
  2. Right-click the chart and click  Show the View Configuration
  3. In the  Properties, expand the Custom Series section
  4. Click the  icon to create a new Custom Series
    1. Name: Enter a name 
    2. Right Axis: Enable this to display the custom series for the data on the right axis
    3. Mode: Choose Percentile
    4. Percentile: Enter a number between 1 and 100
    5. Colour: Choose a colour for the chart line from the colour picker
    6. Style: Select a line style from dashed, dotted or solid

Trend Lines

Trend lines represent the general direction in which data points are moving or trending over time. They help visualise patterns in data and can help make predictions about whether data will be increase or decrease over time. The trend lines available in PhixFlow are: Linear, Quadratic, Cubic and Gaussian.

Below is a bar chart showing how the quantity of different objects change month-on-month. The dashed pink line represents a Gaussian trend. Click on the image to enlarge it.

Creating Trend Lines

  1. Create a line or bar chart, see Creating a Line Chart or Creating a Bar Chart
  2. Right-click the chart and click  Show the View Configuration
  3. In the  Properties, expand the Custom Series section
  4. Click the  icon to create a new Custom Series
    1. Name: Enter a name 
    2. Right Axis: Enable this to display the custom series for the data on the right axis
    3. Mode: Choose Trend Line
    4. Trend Line: Choose from:

      1. Linear - represent the relationship between variables as a straight line.

        This is useful where data increases or decreases along a straight line, at a constant rate.

        Quadratic - represent the relationship between variables as one curve in a line.

        This is useful where data fluctuates up then down, or vice versa.

        Cubic - represent the relationship between variables as two curves in a line.

        This is useful where data fluctuates up and down.

        Gaussian - represent the relationship between variables as a bell-shaped distribution.

        This trend line is also known as normal distribution and is characterised by its symmetric shape around the mean.

        This is useful for showing a symmetrical plot of data around its mean value, where the width of the curve is defined by the standard deviation.

    1. Colour: Choose a colour for the chart line from the colour picker
    2. Style: Select a line style from dashed, dotted or solid

Expression

Create bespoke chart lines using the Expression editor, which has access to the data in each recordset as a data array. Expressions can be used for simple expressions, e.g. 10, or more complex expressions, such as rescaling and rolling averages. The expression should return either:

  • a single value (fixed horizontal line)
  • an array of 2 values (first and last)
  • or, an array of values matching the number of x values/categories

Below is a line chart showing how the value of different objects changes year-on-year. The expression, represented by the dotted blue line, converts negative values for the attribute, Object B, to 0 and doubles the positive values. Click on the image to enlarge it.

Creating Expression Lines

  1. Create a line or bar chart, see Creating a Line Chart or Creating a Bar Chart
  2. Right-click the chart and click  Show the View Configuration
  3. In the  Properties, expand the Custom Series section
  4. Click the  icon to create a new Custom Series
    1. Name: Enter a name 
    2. Right Axis: Enable this to display the custom series for the data on the right axis
    3. Mode: Choose Expression
    4. Expression: Enter an expression using Functions and/or Internal Variables
      This field allows values, e.g. 10, or expressions for calculating rescaling and rolling averages. The expression should return either a single value (fixed horizontal line), an array of 2 values (first and last), or an array of values matching the number of x values/categories.

      The following Internal Variables can be used:

      _value returns a record of arrays

      Example
      //for the attribute, count, convert negative values to 0 and double the positive values

forEach(
		$item, _value.count, 
			if(
				$item <= 0, 0, $item*2
			)	
)

      _this returns an array of records

      Example
      //increases each value for Company A by 10

_this.CompanyA + 10

      Refer to attributes using the syntax, _value.attributeName

    1. Colour: Choose a colour for the chart line from the colour picker
    2. Style: Select a line style from dashed, dotted or solid

More Information