Overview

Dundas Chart for SharePoint™ comes with enhanced formula support for use in financial applications. There are 32 Technical Analyses formulas including Different Moving Averages, Price Indicators, Volume Indicators, Oscillators, and others. In addition to this, there are also Time Series and Forecasting formulas.

Technical Analysis is rooted in Charles Dow Theory. Dow created the Industrial Average, and the Transport Average. He believed that the behavior of the averages reflected the hopes and fears of the entire market. This theory is based on the hypothesis that prices have a trending nature.

Time Series and Forecasting formulas are used to identify the nature of the sequence of observations, and to predicting future values using historical observations (forecasting). In time series analysis, it is assumed that the data consist of a systematic pattern and random noise that usually makes the pattern difficult to identify.

 

The most common chart types used in Technical analysis are:

• Stock Charts (High, Low, Open Close).
• Column Charts (Volume).
• Line Charts (Different Technical Indicators).
• Range Charts (Bollinger Bands, Deviation Chart, etc.).

Applying a Formula an Overview

To apply a formula perform the following:

1. Determine if the data must first be filtered and/or grouped, and then perform the necessary filtering/grouping if required.

2. For averaging formulas only, set the DataManipulator object's StartFromFirst property to the desired value (or use the default).

3. Use the FormulaFinancial method of the DataManipulator class to apply the formula. Note: all formulas have specific requirements (e.g. required input, etc.), and each formula is implemented differently (e.g. default values, etc.). For a general discussion of formula-specific requirements see the section further down. For a detailed explanation of a particular formula, and it's requirements see the Formula Reference folder.

Applying a Formula - Individual Steps

1. Grouping, Filtering and Formulas

In some cases data will have to be manipulated (i.e. filtered and/or grouped) before using a formula.

For example, let's assume that there is raw Stock data with different prices for each day for the last 4 months. The goal is to find a William's %R indicator for this data. We assume the raw stock data and noise are represented by a point chart, and that they were added at design-time. Further, we assume that the DundasBlue template has been applied to the original chart for appearance purposes.

 

Figure 1: Raw data.

 

The first step is to remove 'noise' from the data and to filter all Y values below 40 and above 50. This step is important if we want to remove all anomalous data that incorrectly influences the trend.

 

To do this we will use the Filter method of the DataManipulator class as shown:

 

Figure 2: The Raw data with 'noise' eliminated.

 

The next step is to create a stock chart with Hi, Low, Open and Close prices. First the data must be grouped so that we have the largest, smallest, opening and closing values for each date.

 

This is done using the Group method of the DataManipulator class as shown:

 

The following chart image was created when the code above was used; 'Series1' refers to the original raw filtered data, and 'Series2' refers to the the new stock type series created.

 

Figure 2: Filtered and grouped stock chart that displays High, Low, Open, and Close values.

 

Now we are ready to use the FormulaFinancial method to calculate the William's %R formula as shown:

 

Figure 3: Williams %R formula Chart.

 

Notice that the Williams %R formula doesn't start at the same position where the stock chart starts (Figure 2). This is because the Williams %R formula uses a default period of 14 historical data points, and as a result the first 14 data points from the stock chart will be used to calculate the Williams %R formula. These first 14 data points are not displayed on the chart which presents Williams %R output data.

2. StartFromFirst Property (only applies to averaging formulas)

Since averages of a specified period cannot be calculated for the first number of n data points, where n is the "Period", this property determines how these first data points are dealt with.

Let's assume that we are using a moving average period of 20 days. If StartFromFirst is False (the default), the first 20 data points are just ignored and the moving average will start after 20 days using the previous 20 periods (data points) as historical data only. If StartFromFirst is True, the first 20 data points will be calculated as the n period moving average, where n is current index of data points. (e.g. the third data point will have a value which is equal to the 3 days moving average, while all data points after 20th data point will have 20 days moving average).

3. FormulaFinancial Method

All financial formulas are calculated using the FormulaFinancialmethod, which accepts the following types of arguments: a formula name; input value(s); output value(s), and parameter(s) that are specific to the type of formula being applied.

 

This method has up to four parameters:

1. Formula Name: A FinancialFormula enumeration value that specifies the formula name. This parameter is not optional.

   Examples:

   - FinancialFormula.RateOfChange

   - FinancialFormula.WeightedMovingAverage

2. Input Values: String or Series object(s) which is/are used to specify input Y values and series (this parameter is not optional).
   If the string type is used, input values could be used from different data series and different Y values. If there is more than one input value, then the values should be comma separated. Multiple input values can be used only if the string parameter type is used, and the Series name and Y value index is separated with a semicolon (:). If the Y value is not specified then the first Y value will be used for the formula (the series name must always be specified, however).
   
   Examples:
   - "Series1:Y" - The first Y value is used from series 'Series1'.
   - "Series1:Y,Series4" - The first Y values are used from series 'Series1' and 'Series4'.
   - "Series1:Y,Series2:Y2" - The first Y value is used from series 'Series1'. The Second Y value is used from series 'Series2'.
   - "Series1:Y,Series2:Y2,Series2:Y3" - The first Y value is used from series 'Series1'. The Second Y value is used from series 'Series2'. The Third Y value is used from series 'Series2'.
   
   If a Series object is used for the input value then the first Y value of the series' data points will be used by the formula.

3. Output Values (optional): String or Series object(s) which is/are used to specify output Y values and series.
   Output Values are used the same way as Input Values, the only difference being that output values are optional.
   
   If output values are not specified then the input values will be used to store the formula data, with the first output value stored in the first listed input value, etc. If the number of input values is less than the number of output values an exception will be thrown.
   
   If an output series is specified as a Series object, and the series has not been created and added to the SeriesCollection when the function call is made an exception will be thrown. However, if the series has not been created and added to the SeriesCollection BUT has been specified as a string expression then it will automatically be created at run-time, and have default Series properties (e.g. column chart type, etc.). In addition, the ChartArea property of this series is set to "Default", and if a chart area with the name "Default" exists in the ChartAreasCollection it will be used to draw the chart. If there is no chart area with this name the first ChartArea object in the collection will be used to display the series.

    

4. Parameters: String expression that is used for various formula-dependent parameters. Mostly parameters are used to specify a "period", which is a time period used to calculate a moving average. If a formula needs more than one input parameters (e.g. a short and long period), they should be comma-separated. Some parameters can be omitted and some cannot, depending on the formula type. If a parameter is missing and is not optional an exception will be thrown.
   Examples:
   - "10,20" - Short and long period for Chaikin Oscillator formula.
   - "2,40,true,true" - Polynomial degree, forecasting period, and two booleans that determine if approximation errors and forecasting errors (respectively) are returned.

   
   

   Note
   Many formulas which used a period, use historical data with more data points than the period property. If there are not enough historical data points an exception will be thrown.

    

   Some formulas, especially those that use historical data, will result in output series that have fewer data points compared to the input series (e.g. a 20-day Moving average will not have data points for the fist 20 days). If the same chart area is used to display both input and output series then it may be desirable to filter out all historical data points (data points without corresponding output points) from the input series, after applying the formula method.

Formula-Specific Requirements

The majority of technical analyses formulas use stock prices for calculations. Some of them use Hi, Low and Close values while others use some sort of different prices or values (e.g. volume, etc.). Note: unlike stock charts, which use predetermined Y values (High - Y, Low - Y2, Open - Y3 and Close - Y4), Technical Analyses formulas must always have their Y values explicitly set.

 

As mentioned before, some formulas (Moving Averages) use the StartFromFirst property from DataManipulator class. Some formulas also use default values for parameters, and still others throw an exception if a value is not supplied.

 

 

See Also

 

 

---

Copyright © 2005 - 2009, Dundas Data Visualization, Inc. and others. All Rights Reserved.