EstimationInterfaceReference

The estimation interface is built to experiment with. Unlike most estimation software, TaskEstimation is not built with a single estimation approach in mind. Instead, it provides a set of basic tools and provides what is required to reduce the tedious tasks of data filtering and preparation you would find in a spreadsheet application.

This page intends to explain the interface in details and how it operates. For use cases and detailed steps, see the howto section.

Overview

The interface is divided in four main sections:

• The metric area below the task name. This area allows to add and change the task metrics. These are the same values seen from the task list.
• The data panels on the left side of the screen. The tabbed interface allows to access the task breakdown information and the historical data.
• The short term memory on the top right side of the screen. The box allows to store temporary information between calculations.
• The tool panels on the bottom right of the screen. The tabbed interface allows to access different calculation tools focused on dataset manipulation.

Metric Area

The metric area simply provides a list of all metrics currently available on the task. More metrics can be added by using the text field close to the new icon.

Metric values can be dragged to other fields expecting numeric values. The target field will highlight on mouse over.

Metric values can be modified by clicking on them or by dropping a numeric value on them.

Data Panels

The data panels allow to generate datasets by filtering the elements in the list. The datasets built can then be dragged to other components expecting datasets as an input. Two tabs are available:

• Task Breakdown, allowing to view the details of the currently selected task.
• Data Pool, allowing to access the historical data from previous tasks.

When dragging the title of a numeric column, a dataset is created. The dataset only contains the selected rows — those highlighted in green. One could go through the list and select or deselect rows by clicking on them. However, tags can be used to filter the dataset much faster. Depending on the situation, different techniques may be required. Two fields are available to filter using tags: With and Without. Both can contain multiple tags. In the first case, all tags will be required for the field to be selected. In the second case, none of the tags must be present for the row to be selected. If both are filled, both constraints must apply.

In the task breakdown, parent node of the subtask definition can be unselected automatically by checking the "leaf nodes only" checkbox.

Short Term Memory

The short term memory is a simple but effective tool to hold values between calculations, wether it is to hold your initial dataset or just while switching to a different calculation tool. Any numeric value or dataset can be copied to it either by double-clicking the value or dragging it to the zone. Values can then be dragged out of it.

Values can be removed from it by using Ctrl-Click on the value or by holding Ctrl while dragging it somewhere else.

Tool Panel

The tool panel provides basic mathematical tools that can be used to produce estimates. All of them are aimed at manipulating datasets and to do it in an efficient way. The tools are generic and can be used for multiple purposes. The primary use is to generate time estimates, but they could just as well be used to produce other results provided the right input.

Dataset containers are used in all of them. These containers hold multiple numeric values. When a dataset is dropped into them, all the values it contain can be observed. Multiple datasets can be added sequentially. Single values can also be added to it.

The containers can be emptied by using Ctrl-Click on the individual values, or on the heading for a complete clear. The content of the dataset can be copied to the short term memory by double-clicking the header or dragging it.

Other value fields may be available in certain tools to display the results.

Distribution

The distribution analysis tool takes a dataset as an input and produces basic properties. Produced values include:

• Mean
• Count
• Sum
• Standard deviation

Regression

The linear regression tool is the primary method for estimation once a certain amount of data is reached. To function, the linear regression takes a dependant dataset and an independant dataset of the same length as an input. A minimum of 3 values is required.

The following values are produced:

• R² as an indicator of the data correlation. Values lower than 0.75 for the purpose of software estimation indicate a poor dataset and should be discarded.
• Slope (m) and Intercept (b) as part of your classic y=mx+b
• A graphic to represent the dataset values, the linear regression line and the confidence range.

Changing the Predictor value by either dragging a numeric value on it or editing it directly will change the Forecast and Range values.

• The forecast value is the direct result of the linear regression and can be considered as the most likely outcome.
• The range is calculated through statistical analysis based on the Confidence parameter. The value of this parameter can be changed to values between 0.5 and 0.95. Lower values will provide a smaller but less reliable range.

Calculation

The calculation tool provides a set of very basic operations to open up more possibilities. A drop list allows to choose the calculation to be performed. Upon selection, the required datasets will appear. Some may require multiple datasets. The calculation will be performed for every entry in the first dataset. If insufficient values are provided in the following datasets, the values will loop. Extra values will be discarded.

Available operations are:

• Divide
• Multiply
• Add
• Substract
• Modulo
• Power
• Logarithm
• Square root

---