Portfolio Analysis Options
When you run a Portfolio Analysis you can optionally choose to first specify options within the Portfolio Analysis Options dialog. If you'd like to specify these options prior to a run, drop-down the Portfolio | Run | Analyze split button and choose Specify options and analyze... from the list. You can simply click the top half of the split button if this option is indicated by the icon (i.e., it includes a gear). The Portfolio Analysis Options dialog has seven sections; three of which correspond to input settings for the run and four of which correspond to output options from the run.
If you have defined categories for the projects in your portfolio, you may choose to run a subset of them by project type. If you wish to run all projects in your portfolio, set the radio button to All projects. To run a subset of projects:
- Set the radio button to Subset.
- Click Select. The Project Selection dialog appears.
- Select the project types you wish to run. Specify whether or not to include projects with missing data by checking the box beneath each category type.
- Click OK to close the Project Selection dialog.
Whether you've defined categories or not, you may also filter projects via the Portfolio Filter dialog by choosing the Edit... button to the right of the Filter checkbox. If you've already defined a filter, check the Filter check box to apply it.
If you run a subset of projects and filter, then the projects included in the run will be those that are in the subset and meet the filter criteria.
This section only applies to database-linked portfolios.
You may choose to run your portfolio with latest data that is extracted from the database at run time or with previously extracted data. The "Use previously extracted" option is only available if database data has been previously extracted.
If you choose to run by extracting data from the database, you will need to have access to the database. See Database Overview for more information on connecting to a database. To change the database data extraction options, click Settings. The Database Data Extraction Options dialog appears.
If you know you will be running an analysis remotely (i.e., when you do not have access to the database), you should extract the data while you have access to the database (Portfolio | Project Data | Extract) and save your Workspace file. Then, you may run with the "Use previously extracted" option when you are remotely located. Further, extracting the database data does take time. If you will be doing a number of runs, you may wish to extract the data the first time and create data files from the data (see Data Extraction Options). Subsequently, you can use the previously extracted data, which is stored in DPL Programs in the Workspace.Revisions
This functionality is only available if your database is configured to have revision tracking. See Database Overview for more information on configuring your database for revision tracking.
If the database has revision tracking, you may choose to run a previous revision of the data (i.e., a revision other than the most current). You may do this by telling DPL to use a specific Revision ID or by choosing a date and time for the revision. If you run a previous revision, DPL will select the data that is the revision that is as recent as possible but earlier than the Revision ID or revision date and time. E.g., if you ask for a revision date and time of March 1, 2008 at 10:00am (and assuming there is no revision at that exact time), then a revision with a date and time of March 1, 2008 at 9:59am qualifies but not one at 10:01am. If there is no revision earlier than the specified revision, DPL will choose the earliest revision of the data.
If you would like to see a grid view of the portfolio results, check "Results grid". A grid view will be produced with the expected values and 10th, 50th and 90th percentiles of the objective function and each attribute of the portfolio. The projects are in rows and the objective function and attributes are in columns. Four sheets are produced one each for the expected values and three percentiles.
In addition, if your portfolio is linked to a database, you can export the expected values to a results table. See Database Specification and Portfolio Attributes for more information on exporting expected value results to a database.
If you would like to see one or more Risk Profile Charts for the entire portfolio, check the box labeled Portfolio Risk Profile Chart and use the drop-down box beneath it to specify the charts to produce. If you'd like DPL to produce risk profile datasets for each individual project in the portfolio, check the box labeled Project Risk Profile Data and use the drop-down box beneath it to specify which datasets to produce. Risk Profile charts for individual projects can be subsequently created using this data. See Risk Profile data. A label at the bottom of this section indicates how many charts and/or datasets you have selected to be produced.
DPL offers six different views for comparing the projects within a portfolio as summarized below.
Risk/Reward Bubble Chart: check this box to produce a bubble chart that plots the probability of development success on the y-axis vs. the expected objective function value on the x-axis. The size of the bubble is relative expected development cost.
Project Value Range Chart: check this box to produce a bar chart that shows the 10th to 90th percentile range of value (and expected value) for each project in the portfolio.
Productivity Ratio Chart: check this box to produce a chart that plots the expected objective function value on the y-axis vs. the expected development costs on the x-axis. The plot ranks the projects from high to low of the ratio of expected objective function value over expected development costs.
Portfolio Tornado: check this box to produce a tornado chart that shows how each individual project contributes to the overall uncertainty of the portfolio.
Expected and Potential Value Chart: check this box to produce a column chart that shows expected objective function value and the expected value given development success for each project in the portfolio.
Initial Decision Alternatives Productivity Chart: check this box to produce a chart similar to the Productivity Ratio Chart except that initial decision alternatives are included for each project.
Time Series Percentiles
DPL will produce percentile charts for the entire portfolio or for individual projects. Check the appropriate box. If either is checked, indicate which time periods to calculate the percentiles from and to. Also provide a label for the initial time period.
Versions: DPL Portfolio