Welcome to the Population Analyst, an interactive tool for visualizing population structures and creating population projections.
Population visualization is achieved through the creation of population pyramids, which are graphs of the age and sex distribution of a give population (Weeks, Population, 1999 p. 279). They provide a visual depiction of a population’s composition; this graphic representation offers immediate clues about a population’s demoraphic characteristics. Their fundamental purpose, then, is to reveal demographic patterns that would not be seen or understood as easily when looking at tables of raw data.
Population projections are conducted using the cohort-component method, which uses specific information about each age/sex cohort to estimate its future population. The Population Analyst’s implementation of this technique employs three fundamental components of population change: fertility, mortality, and migration. The accuracy of a projection is obviously limited by the accuracy of its parameters.
The Population Analyst operates on population datasets. A population dataset enumerates the size of each cohort in a population; a “cohort” is a set of people of a particular sex whose ages fall within a particular range. The Population Analyst uses five-year age intervals starting with 0 to 4 and ending with 85 and up, so there are 18 different age groups yielding a total of 36 cohorts (18 age groups × 2 sexes = 36 cohorts). Each dataset includes three identifying fields: name, year, and source.
It is the varation among these cohorts that population pyramids vividly illustrate. As the name implies, cohort-component projections project each cohort’s population individually, based on its specific fertility, mortality, and migration rates.
The Population Analyst offers a number of functions for managing these population datasets. These functions are described in the following sections.
The Population Analyst has two parts. The green strip on the left is the Datasets panel, which lists all the datasets you've made available for analysis and provides the controls necessary to initiate each function. The wider section on the right displays the controls and results for each particular function.
When you first load the Population Analyst, the Datasets panel will resemble the image at left. This initial screen contains no controls except for a button to load some data and a link to display the introductory welcome page (useful if you want to get back to the beginning after clearing all the datasets you’ve been working with).
After loading a dataset, the Datasets panel will change to resemble the labeled image at right. The top row contains buttons to apply the Population Analyst’s primary functions to the selected datasets. 
These two buttons are followed by a menu containing the various data management functions (shown in the adjacent image); the Go button carries out the selected function.
These controls are followed by a list of all loaded population datasets. Each dataset is preceded by a checkbox, which you can clear to temporarily omit that dataset from whatever operation you intend to carry out. Beneath this list is a row of selection buttons which you can use to quickly select or deselect all datasets (useful if you have loaded many). The Switch buttons inverts the selection so that unselected datasets are selected and selected datasets and unselected.
Using the Load Data function will initially display a page labeled “Load Data From Where?” This page offers three options, shown at right and listed below.
The first option will display a list menu similar to the one shown at left. This is a list of every place and corresponding population dataset provided for the chosen geography. You can make multiple selections from this menu by holding down the Shift or Control buttons while clicking on different entries. You can load as many datasets as you want (useful for comparison purposes), but be warned that the more datasets you load the slower the Population Analyst may perform.
The provided datasets are obtained from reputable sources such as the U.S. Census Bureau. Their associated projection parameters are derived from state health statistics and additional Census Bureau data.
The existing data file option will display an upload prompt similar to the one shown at right. You can use the Browse button to select a data file residing on your computer. This file may be one you have previously saved or one you have just prepared from a different data source.
Population Analyst data files are plain text files. The first three lines identify the name, year, and source of the dataset.
These three lines are followed by 36 more, each containing a single number. The first 18 represent male cohort populations and the second 18 represent female cohort populations. Each cohort is a gender-specific age group defined by five-year intervals, as described in the Data Management section.
Data files saved from the provided datasets may contain additional lines referencing the location of the associated projection parameters. These lines are not strictly required, but must be retained for the Population Analyst to recognize which fertility, mortality, and migration rates should be associated with that data.
The last input option will display a form like the one pictured (at reduced size) to the right. You can use the form to create a new population dataset from scratch, which can then be saved to a file for use again later.
After you’ve selected, uploaded, or entered your new datasets, the Population Analyst will present a brief confirmation page asking for your approval. When you grant it by clicking the Finish Loading button, the new datasets will be appended to the list in the Datasets panel.
You can use the Save Data tool to save population datasets to your computer as plain text files. The function will print a list of links to the datafile for each selected entry. Clicking these links will load the text file directly in your browser, which you may then copy or save. Another way is to right-click on the link and choose the Save As option from the menu that appears, as shown at right (the Save As option may be named slightly differently in different browsers, but it will usually be recognizable).
Saved datasets may be edited with a text editor or, with modest effort, imported into a spreadsheet. They can also be uploaded at a later date using the “Existing data file” loading option discussed above.
If you’ve made an error in your manually entered data or if you would like to experiment with any other changes to the datasets you’ve loaded, you can use the Edit Data tool to modify your datasets. The interface is very similar to the manual data entry form, except the fields are automatically filled in with their current values.
A confirmation page similar to the Load Data confirmation will be displayed after you submit your changes.
Changes made using the Edit Data function only affect the temporary instances of the data contained in the Datasets panel. The original provided data is not changed nor are any files modified on your computer. Therefore, if you would like your changes to be permanent you should save the modified dataset after confirming your edits.
Use the Delete Data tool to remove datasets from the Datasets list. Only the checked datasets will be removed.
As with the Edit Data function, deletion only removes the instances of the data present in your Datasets panel. The original provided data is not effected nor are any files removed from your computer.
The Examine Data function displays a table containing all the selected datasets' cohort populations. This is useful for looking up the specific figures corresponding to interesting features on a population pyramid.
In modern browsers like Firefox, the row beneath the cursor will be highlighted a darker shade of green (as shown in the image at right) to facilitate comparison of the same cohort across multiple datasets. Viewing the entire table may require horizontal scrolling if many datasets are selected.
Clicking the column heading links in the above table will open a new window with a similar table listing the selected dataset’s mortality, migration, and fertility rates. These are only defined for the provided data; other datasets will be assigned default survival rates of 1 (no mortality) and migration and fertility rates of 0. These projection settings introduce no changes besides a shift towards older age groups. Projecting unique datasets therefore requires corresponding projection parameter files to be supplied prior to projection, as discussed in the relevant section.
The column heading links in the projection parameter window are similar to the Save Data links in that they point to plain text versions of the corresponding data. You can save these files to your computer using the same technique, modify them while retaining the same format, and supply them to the projection function as parameters for your custom datasets.
Click the Pyramid button in the Datasets panel to start creating population pyramids for the selected datasets.
A configuration page will appear, offering you many options for controlling the structure and appearance of your population pyramids. The Pyramid Templates section at the top of the page offers a few basic designs to get you started. Clicking a template will automatically set the options necessary to recreate that same style. You can then click Continue to go with that design, or you can adjust the options yourself. The options are set by default to create a pyramid resembling the first template.
After clicking Continue, the selected datasets will be graphed. You can save the resulting population pyramid images by right-clicking on them and choosing the Save Image option, as shown at right. Some browsers may also allow you to copy the image or drag it directly to a folder.
Click the Redesign button on the pyramid page to return to the configuration page. Many of the configuration options are self-explanatory, and the rest are explained by contextual help boxes.
Certain option and dataset combinations may produce curious pyramids like those shown at left. These pyramids were drawn with a common scale based on the actual populations of a small rural county and a very populous urban county. 
This configuration effectively shows the differences in magnitude between these populations, but it obscures the structure of the smaller population.
Redesigning the pyramids to use percent scaling will normalize the size of the pyramids, thereby emphasizing their actual structure. The second revision, shown at right, clarifies the structural differences between these rural and urban populations.
This is just one example of the analytical power of population pyramids.
Click the Projection button in the Datasets panel tp display the projection configuration page, which resembles the image below.
The minimum projection interval of 5 years is recommended as the most accurate and informative. The same initial projection parameters are used for each longer interval, meaning the more distant the projection the less applicable and therefore less accurate the parameters will be. The recommended strategy for conducting long-range projections is to project 5 years at a time and upload custom parameter files modified for the projected circumstances at each interval.
Deselecting the “Include migration” option allows you to discount the (often significant) effects of migration, thereby achieving a “natural increase” projection. Again, this option is offered for experimentation but the default checked status is recommended for the most meaningful projections.
The "Examine" link displays the same projection parameter table described in the Examine Data section. Clicking the "Upload Custom" link will reveal three fields for uploading custom survival, migration, and fertility rate files, as shown at right. The uploaded parameter files must be in the same format as those downloadable from the parameter table window.
Uploaded parameters will override a dataset's default associated parameters, if any. All three parameters files do not have to be uploaded; those uploaded will be used, and the default parameters will be used for the others.
Like the Load and Edit Data functions, the projection function will present a confirmation screen prior to adding the projected datasets to your Datasets list. Click the "Add Projected Datasets" button to update your list of datasets.