Difference between revisions of "Data Analysis"

From Wiki
Jump to: navigation, search
 
(19 intermediate revisions by 2 users not shown)
Line 1: Line 1:
 
= <span style="font-size:xx-large;">World Map: Historic Data</span> =
 
= <span style="font-size:xx-large;">World Map: Historic Data</span> =
  
The World Map option is a constantly evolving program, with new features added to the option regularly. The World Map is located under the Data Analysis option on the Main Menu.
+
The World Map is located under the Data Analysis option on the Main Menu of IFs.
  
The main function of the World Map is to display data in a visually dramatic format. Instead of representing country-related forecasts as lines in a&nbsp;'''graph''', or columns in a&nbsp;'''table''', the user is able to view the forecasts as shades of color imposed on a map of the world.
+
[[File:Worldmap1.png|frame|center|World Map example]]
  
A function of the menu of the map feature is&nbsp;'''Continue'''&nbsp;button. Use this to exit and return to the previous display screen.
+
The main function of the World Map is to display data in a visually dramatic format. Instead of representing country-related forecasts as lines in a graph, or columns in a table, the user is able to view the forecasts as shades of color imposed on a map of the world.
  
Another menu option is&nbsp;'''Display Data'''. If this option is selected, a box will appear that allows users to change the data set that is being displayed. The default data set that is available is the TimeSeries data table. By scrolling through the Data Table, users can select from different data sets. These sets then correspond with a larger list of variables displayed in the Variable Name scroll-down. Below the Variable Name option is the Dimension of Variable which allows users to change the year being displayed. At the top right of this box is the ability to increase the amount of categories displayed as well as switching between Equal Interval or Equal Count. The later option allows users to either display an equal number of countries in one category or numerically equal categories. Finally, at the very bottom of this box, users are able to access the Data Information.
+
The main features of this option are located at the bottom of the screen which allows users to change the data set that is being displayed.&nbsp;'''Be sure to click the Refresh Map button after making any changes.'''
  
Another feature of this version of the map is Display Options. The following options can be found by selecting this option.
+
*Use&nbsp;'''Continue'''&nbsp;button to exit and return to the IFs Main Menu.
 +
*Data availability statistics and source information of each data series are available from the&nbsp;'''Data Information'''&nbsp;button.
 +
*The default data table that is available is TimeSeries, a broad group that contains most longitudinal data series in IFs. By clicking the&nbsp;'''Data Table'''&nbsp;drop-down menu, users can select from different data sets. These sets then correspond with a larger list of variables displayed in the&nbsp;'''Data Field'''&nbsp;list box.
 +
*Below the Data Table option is the&nbsp;'''Dimension of Variable'''&nbsp;which allows users to change the year being displayed. Options include individual years, earliest year available for each country and most recent year available for each country.
 +
*Boxes labeled&nbsp;'''Display Type'''&nbsp;and&nbsp;'''Number of Categories'''&nbsp;allow users to customize the map display. The default display type is set to Equal Interval, which divides the total range of values into the number of categories specified. The alternative categorization, Equal Count, displays an equal amount of responses in each category. For any data display option, users can create up to 7 distinct categories for display.
 +
*Another input box,&nbsp;'''Color''', determines the shading color in which data is displayed.
 +
*Finally, by selecting the&nbsp;'''Projection'''&nbsp;option at the very bottom of the page, users are able to toggle between multiple global cartographic projections.
  
*'''Labels:'''&nbsp;Selecting this option will allow users to place the names of all countries on the map. Removing Labels will take these names away. Users can also change the font of the country names.
+
Some final features of the map are located inside the map display.
*'''Colors:'''&nbsp;By selecting this option, users have the ability to change the color in which data is displayed.
+
*'''Projection:'''&nbsp;By selecting this option, users have the ability to display the map differently.
+
*'''Layers:'''&nbsp;Selecting this option presents users with the ability to add different layers on top of the map to more clearly see how geographic changes may or may not play a role in different data displays. The layers that the user can add include infrastructure, such as roads, railways, and piplines, and environmental, including options such as rivers and forest cover. The user is also able to remove all or specified layers.
+
  
Another feature of the map is the collection of different buttons located directly beneath menu options.
+
*The map legend is located to the left of the map. It displays the data points that delineate the different display categories.
 +
*To move around the map, use the arrow buttons located to the right of the map; to reset the view, click the button at the center of the arrow buttons; zoom in or out using the plus/minus bar.
 +
*Hovering over each country in the map displays the associated value of the data set, and left clicking on the map allows users to view either a detailed map of the country or a table of the data values by year.
  
*'''A magnifying glass next to a plus sign:'''&nbsp;can be used to zoom in on a specific part of the world. Simply highlight the button, move your mouse over the part of the world you would like to more closely look at and left click.
+
= <span style="font-size:xx-large;">Analyze Across Countries (Cross-Sectional Analysis)</span> =
*'''A magnifying glass next to a minus sign:'''&nbsp;can be used to zoom out. Simply highlight this button, move your mouse over the world map and left click.
+
*'''A hand:'''&nbsp;can be used to move the world map around. Highlight this feature, move your mouse over a zoomed-in world map, left click, hold and drag to your desired location.
+
*'''A globe:'''&nbsp;can be used to zoom back out to the standard, centered, default view of the world map.
+
*'''A printer:'''&nbsp;can be used to print out the map you are looking at.
+
*'''A pointer:'''&nbsp;can be used to copy the data that underlies the map display. Clicking on this button will present a prompt that tells you that you have copied the data to the computerclipboard. If you would like to, say, copy the data to a Word file, simply open the file, right click on the white, empty space and select paste.
+
  
Some final features of the map are located to the left of the map and below the map. The map legend is located to the left of the map. It displays the data points that delineate the different display categories. On the bottom of the map display, the year being displayed is identified as well as the variable name.
+
From the Main Menu, an option under Data Analysis provides users with the ability to examine the relationship between historical variables. From the Data Analysis menu, select Analyze Across Countries (Cross-Sectional) to reach the screen displayed below.
  
= World Map Movie =
+
[[File:Crosssectional1.png|frame|right|Variables across countries window]]
  
The&nbsp;'''World Map Movie'''&nbsp;is accessible in two locations in IFs. The first is under the&nbsp;'''Data Analysis'''&nbsp;heading on the main IFs screen. The second is under Specialized Display, which is a sub-heading of the Display heading on the Main Menu. The World Map Movie found under the Data Analysis heading deals with historic data, while the World Map Movie found under the Specialized Display heading deals with forecasted data. Selecting the World Map Movie allows the user to display on a map of the world the changes in selected variables over time in all of the countries for which IFs has data. This section describes how to use the various options to tailor the movie to the userpreferences.
+
To do a cross-sectional analysis, specify a dependent variable (a variable you want to understand the causes of) and one or more independent variables (the possible causes of change in the dependent variable). By convention, the dependent variable appears on the y-axis and the independent variable appears on the x-axis on a graph. To select a dependent variable first, scroll through the alphabetized variable list and left-click on any variable in which a user might be interested. Click Select to choose that variable. Users can also type a variable name in the simple text search box directly above the list, which filters series by prefix. Alternatively, there is a more sophisticated search feature available through the Search button at the bottom of the screen. From the search screen, a search on "female", for example, would find all results that contain the word, anywhere in the variable name or description. After selecting a dependent variable, either from the variable list or from the search form, a drop-down list will appear, specifying the year to be displayed. A cross-sectional display includes data for a single year, earliest available or most recent available for each country. The process described here is also used to assign independent variables. To reassign any variable, select the corresponding option button and simply load the new series.
  
After selecting World Map Movie, the dialog box pictured below appears on the userscreen.
+
As a demonstration, plot a cross-sectional relationship between female education and reproduction rates. For the dependent variable choose total fertility rates, by selecting TFR from the list of variables. Left-click the variable and two options will appear: Select and Data Information. Click Select to choose that variable. The latter provides detailed information about the scope, availability and source of the data set. The ensuing drop-down list determines the year for which data will be displayed. For this example, choose 2010. Notice that TFR(2010) is then shown as the Dependent Variable. Next, choose an Independent Variable, the level of adult female education, by selecting EdYearsAge15Female for the same year. Click the Plot button to generate the graph displayed below.
  
image&nbsp;http://www.du.edu/ifs/help/use/data/movie.html
+
[[File:Crosssectional2.png|frame|center|Scattergram of raw country data example]]
  
'''Choose Variables and Years'''
+
To learn how to use various features on the scatter plot, click&nbsp;[[#Cross-Sectional_Scattergram|here]].
  
*'''Variables''': allows user to select from the entire database of variables in IFs for the Historic World Map Movie. The forecasted World Map Movie includes a more limited database of variables from which to choose.
+
== <span style="font-size:x-large;">Cross-Sectional Scattergram</span> ==
*'''Start Year''': Enter the year from which the movie will begin.
+
*'''End Year''': Enter in the year with which the movie will end.
+
*'''Filter''': User can set the filter from 1% to 100%. This feature sets a minimum value for a portion of the data to be displayed in a given year.
+
*'''Category''': Can be set from 1 to 16. Determines the number of categories to divide the values into.
+
*'''Constant/Variable Legend''': Selecting the constant function will keep the numbers and measurements in the individual categories constant. The numbers and measurements will change over time if the variable legend is selected.
+
*'''Display Type''': When Constant Legend is selected, two options will appear under display type: equal interval and equal count. The equal interval function divides the values equally between categories, with the interval calculated by the equation (highest value-lowest value/number of categories= interval width). The equal count function distributes the number of countries equally between the selected of number of categories. When Variable Legend is selected, a third option becomes available: equal match. If equal match is selected, legend values are assigned to each value uniquely. If there are not enough categories to represent each value, the countries that do not have the selected values are represented as blank.
+
  
'''Controlling Options'''
+
This section describes how to use the various features on the Scattergram, which appears after clicking the Plot button in the Analyze Across Countries (Cross-Sectional) option.
  
*'''Automatic''': When deselected, the user scrolls through the years of the movie manually. When selected, the movie will automatically play the start year to the end year.
+
[[File:Crosssectional3.png|frame|center|Scattergram of raw country data example]]
*'''World/Country''': This toggle switch is available only under the Historic World Map Movie. It allows the user to view a movie of changes of the selected variable across the world or, when country is selected, to view a movie of the change in a variable across regions in a country. Three countries are available to the user to select: China, India, and Mexico.
+
*'''Create Movie file''': This toggle switch allows the user to save the movie created. When the playing of the movie is complete, a prompt appears which allows the user to name and save the movie file. The saved movie file is available in Stored Map Movie.
+
  
After selecting Start, the dialogue box will close, and the World Map as pictured below (or similar to it) will appear. This section describes what the user is viewing, and how to manipulate the functions at the bottom of the screen.
+
The&nbsp;'''Relationships'''&nbsp;option allows users to fit a line to the graph. To impose a line of relationship on the scatter plot as displayed above, choose Relationships from the toolbar, and the option Linear. Notice the r-squared value (a measure of the goodness of fit between the independent and dependent variable) and linear equation displayed at the top of the screen. Additional relationship types include: logarithmic, exponential, power, polynomial and logistic. To remove the line, click on No Regression.
  
image&nbsp;http://www.du.edu/ifs/help/use/data/movie.html
+
The&nbsp;'''Display Labels'''&nbsp;feature allows users to change the labels on the graph. The default selection is No Labels. Although the country names are not shown on the graph with this setting active, users are able to see them by hovering over the data points on the plot area. By switching options under the Display Labels menu, countries can be represented by the full names or the abbreviations of the names instead. Users can also fully customize the labels of the x- and y-axes, the title, and the subtitle of the graph.
  
== Playing Movie for Historic Data ==
+
Alongside Display Labels, the&nbsp;'''Specify Geography'''&nbsp;menu option allows the user to display data by country, group, or geographic list.
  
The upper left portion of the screen lists whether the movie is showing a variable or constant legend and across which period of time. Underneath this information is the number of categories and the corresponding color for each category, and the numerical range for each category.
+
Clicking on&nbsp;'''Continue'''&nbsp;returns the user to the previous page.
  
The main portion of the screen is the world map, divided into all countries for which IFs has data. The various shadings of the countries correspond to the various shading of the categories listed in the map legend. By right-clicking on the world map, the user is able to copy the map, zoom in on a particular location of the map, zoom out, or to reset the map to its original point of view.
+
= <span style="font-size:xx-large;">Analyze Across Time (Longitudinal Analysis)</span> =
  
The bottom section of the screen, from left to right, lists the variable displayed by the map, with the start year shown and the level of the filter. Next is the selected display type. The four buttons to the right of the display type allow the user to zoom in, zoom out, pan, and show the full extent of the world map. The first two buttons allow the user to zoom in on or zoom out from the area under which the cursor is located, while the full extent option resets the view of the world to the original perspective. The pan button allows the user to pan across the world by clicking on different parts of the map, which effectively pans across the world by re-centering the map on the area selected, after having zoomed in or zoomed, without zooming in or out any further. The next set of buttons allows the user to play or pause the movie, and to adjust the speed at which the movie will play. Next is a toggle switch which, when selected, makes the movie play automatically; when the switch is not selected, the user is able to advance the year displayed on the map or to go over previous years.
+
Whereas&nbsp;[[#Analyze_Across_Countries_.28Cross-Sectional_Analysis.29|cross-sectional analysis (Analyze Across Countries)]]&nbsp;concerns historical relationships between different variables, the primary purpose of the longitudinal analysis (Analyze Across Time) is&nbsp;to trace historical trends of individual data series. This option can be accessed from the Data Analysis option on the Main Menu of IFs.
  
If the user selected “Create Movie File” during the setup phase, a dialogue box will appear after the movie is finished playing that prompts the user to provide a name and a place to save the movie file. The program will alert the user if the .avi file is successfully created and saved. To review the saved movie file, go to Stored Map Movie.
+
[[File:Longitudinal1.png|frame|right|Variables across time window]]
  
= Analyze Across Time (Longitudinal Analysis) =
+
To start with, pick any data series to load a dependent variable. Time is set as an independent variable by default. Users can select a variable by scrolling through the dependent variable list box, typing a variable name in the text box or with the more sophisticated, separate Search option, located at the bottom of the screen. Click Select Countries to specify the data that will appear in graphs. (Note that Use Groups , the button beneath Select Countries, is a toggle, switching from country to group selection or vice-versa; e.g. if groups are selected the toggle will read "Use Countries"). From the country selection screen, pick a country, a group or a decomposed group. Buttons on the right side of the country selection screen also allow users to switch aggregation to groups or decomposed groups. Continue back to the original screen for analysis over time and Plot the result.
  
This option can be accessed from the Data Analysis option on IFs Main Menu.
+
To demonstrate, pick a variable such as AidDon%GNI (aid donations as a percent of Gross National Income, the successor concept to Gross National Product). Click the Select Countries button and choose United States as a donor country. To generate the graph displayed below, click Plot. You will see the trend of U.S. aid commitment over time.
  
image&nbsp;http://www.du.edu/ifs/help/use/data/analyzetime.html
+
[[File:Longitudinal2.png|frame|center|Graph of Longitudinal Data Example]]
  
This screen looks much like others you have seen. One difference is that you must select a single or a small set of countries for your longitudinal analysis. You can look at a variable over time (using time as the independent variable), or you can treat time as points on the graph and look at the relationship between two variables.
+
On the Graphical Display screen as shown above, options on the top menu bar allow users to go back to home screen, alter the scale, display a table, format titles and explore future trends. Under the Trend menu, users can impose extrapolations—linear, polynomial, logarithmic, exponential and s-curve—onto the chart. The sub-option, Extrapolation Setup, features customization options for polynomial, exponential and s-curve trend lines.
  
Pick a variable such as AIDDon%GNI (aid donations as a percent of Gross National Income, the successor concept to Gross National Product). Using the screen that comes up when you touch Select Countries, pick the United States as a donor country. Then Exit back to the screen for analysis over time and Plot the result. You will see the decline in U.S. aid commitment over time.
+
An additional feature of the longitudinal analysis is the ability to look more closely at the data set in which a user might be interested. Left-click on any variable and a small window with two options will appear: Select and Data Information. Click Data Information in order to access detailed information as to the scope, availability and properties of the data set in question.
  
On this particular plot, and dependent on the version of the model you are using, you may see values for Earliest and MostRecent. These values have been entered into the data file so that users can have available; for example, the most recent value of a variable for all countries, regardless of whether that value comes from 2000, 2001, or 2005. Most recent (and earliest) values are especially useful for cross-sectional analysis and you may or may not want them on your longitudinal&nbsp;[http://www.du.edu/ifs/help/use/repeated/general/graph/graph.html graph].
+
Some other features of this section of IFs include the&nbsp;[[Repeated_Features#Computations|Computations]]&nbsp;button and the&nbsp;[[Repeated_Features#Table_Use|Table]]&nbsp;button.
  
Another feature of this section of IFs is the&nbsp;[http://www.du.edu/ifs/help/use/repeated/computations/computations.html Computations]&nbsp;button.
+
Other than the default setting, treating time as an independent variable, which allows users to look at a historical trend of one variable over time, two other types of analysis are also available in the model: treatment of time As Points and As Time. When the Treatment of Time check box labeled As Points is active, the data points become years and users can look at the relationship between two variables as x- and y-axes. The third option, treatment of time as time, facilitates side-by-side comparison of longitudinal trends with two different variables on the same display (the x-axis is time, in this case).
  
 
Exploration of maps, cross-sectional and longitudinal relationships, and relationships computed for IFs can give you much information about the apparent relationships among a wide range of global development indicators. Even if you never used IFs for forecasting, this data analysis capacity could significantly enhance your understanding of the world.
 
Exploration of maps, cross-sectional and longitudinal relationships, and relationships computed for IFs can give you much information about the apparent relationships among a wide range of global development indicators. Even if you never used IFs for forecasting, this data analysis capacity could significantly enhance your understanding of the world.
  
= Analyze Across Countries (Cross-Sectional Analysis) =
+
= <span style="font-size:xx-large;">Show Computed Functions</span> =
  
This option can be accessed from the Data Analysis option on IFs Main Menu.
+
To access Show Computed Functions, click on Data Analysis on the Main Menu of IFs.
  
image&nbsp;http://www.du.edu/ifs/help/use/data/crosssectional/index.html
+
By clicking on Show Computed Functions, you will be presented with a menu that looks similar to the one pictured below:
  
The menu of Analyze Across Countries is dominated by the Select Variable and Data Table. From this list, users are able to select from all of the data tables that underpin the functioning of IFs.
+
[[File:Functions.png|frame|center|Show Computed Functions menu]]
  
Note that the Select Variable and Data Table list box includes many variables that do not exist, or do not have the same names, in the IFs model. These are from the data base. When possible, data were gathered for many historic years, in order to provide the ability to analyze across time (longitudinal analysis, described later) as well as across countries at a given point in time (cross-sectional analysis). When you select a variable name for which data from more than one year is available, you will be asked to specify a year.
+
The box that is located in the middle of your screen holds hundreds of functions. If you click on the Extend List option right above this box, it will display an expanded list of functions.
  
To do a cross-sectional analysis, specify a dependent variable (a variable you want to understand causes of) and one or more independent variables (the possible causes of change in the dependent variable). For instance, pick life expectancy (LifExp – SocioPolitical (Life expectancy at birth) - years) as the dependent variable. Many relationships in IFs treat Gross Domestic Product (GDP) per capita at purchasing power parity as an independent variable (GDP2000PCPPP – Economic (GDP per capita (PPP)) – 2000 PPP$). Click GDP2000PCPPP in the Select Variable list box. You will then have to choose a year. For this example, choose the year 2000. Notice that GDP2000PCPPP(2000) is then shown as the Independent 1 Variable. Click the Plot button to access the display of these variables.
+
In the Filter Functions box on the top of your page, type in any search terms for functions that you would like to see displayed and hit the Enter key. This will bring up all the functions/relationships that have your term(s) in their names.
  
An additional feature of the Cross-Sectional Analysis is the ability to look more closely at the data set in which a user might be interested. Left-click on any variable and a small window with two options will appear: Select and Data Information. Click Select to choose that variable. Click Data Information in order to access detailed information as to the scope, availability and properties of the data set in question.
+
After you have selected a function, choose the Graph option from the top of the menu and click on Draw. This will then display the function at the bottom of the screen along with the specific function and an R-Squared value.
  
At the bottom of the dialogue box is a drop-down menu with the title&nbsp;[http://www.du.edu/ifs/help/use/repeated/source.html Source File]. Choosing different source files allows the user to view and experiment with data for a small number of countries at the provincial level.
+
Select one or more of functions if you would like to see multiple relationships on one screen. In order to select more than one function, highlight the first function, hold the Ctrl key and select a second function.
  
== Cross-Sectional Analysis Extened Features ==
+
= <span style="font-size:xx-large;">Identify Group or Country/Region Members</span> =
  
These features can be accessed through&nbsp;[http://www.du.edu/ifs/help/use/data/crosssectional/index.html Cross-Sectional Analysis]&nbsp;from Data Analysis on the Main Menu of IFs.
+
To identify regions or group members used in IFs, click on Data Analysis on the Main Menu of IFs, and then select Identify Group or Country/Region Members.[[File:Identify.png|frame|right|Identify Group or Country/Region Members window]]
  
'''Extended Option.'''&nbsp;When you use the input line for independent or dependent variable selection, it is also possible to compute a variable as a simple function of two or more other variables (e.g. as the sum, difference, quotient, or product of two other variables).
+
This feature of IFs allows users to identify which countries/regions are represented in certain groups. By clicking on any of the groups in the left-hand list, the members of that group will appear on the right hand side.
  
Now touch the Plot button to create a scatter plot showing the relationship between the two selected variables for as many countries as exist in the data set. You will see that life expectancy increases rapidly as GDP per capita rises. The scatter plot may be sufficient for many users of IFs. You can print the scatter plot or save it for other analysis or for use in other applications.
+
By clicking on Using Regions, the list of groups on the left hand side will become a list of countries/regions.
 
+
Some users, however, will want to proceed further and to describe the relationship between the two variables with an equation. If you have Microsoft Excel available on your computer, you can do this by touching the Excel button from the Scatter plot. That button actually creates a link with Excel and carries the scatter plot (with supporting data) to Excel for further analysis.
+
 
+
http://www.du.edu/ifs/help/use/data/crosssectional/crosssectional.html image
+
 
+
Excel opens up many additional features for your use that can be described in more detail using Excel’s Help menu. But let’s walk through the process of fitting a line and associated equation to the data (important for the next Lesson). First, maximize the Excel screen by using the maximize button in the upper right hand corner of the window (the icon of a single large window).
+
 
+
Then click anywhere on the graph to activate it. That allows you then to right-click on various components of the graph in order to edit those components. For example, right-click on the straight line that Excel has fit to the data. You will see a small pop-up menu with a couple of options. Select the Format Trendline option and you will see the following window.
+
 
+
http://www.du.edu/ifs/help/use/data/crosssectional/crosssectional.html image
+
 
+
Choose the Type tab and select the logarithmic trend/regression type for a better fit.
+
 
+
Then choose the Options tab for additional choices including checking boxes for "Display Equation on Chart" and "Display R-Squared Value on Chart." Although both options are checked, the equation and R-Squared are sometimes not initially visible on the chart because they are not placed properly when you open Excel (it depends on the variables you plot). If they are not visible, turn the options both off, close the Format Trendline window, right-click on the line, select again the Format Trendline option, and again select the Options folder. Turn the Display Equation and Display R-Squared options back on and exit from the Format Trendline window again. You should now see the equation and r-squared someplace on the graph and can drag them to a better location, nearer the intersection of the two axis.
+
 
+
If you now right-click on the dependent or y-axis, you will see a small pop-up menu with several options. Choose the format axis option and the following window will appear:
+
 
+
http://www.du.edu/ifs/help/use/data/crosssectional/crosssectional.html image
+
 
+
Under the Scale tab you will find maximum value, minimum value, and other settings. Often you will find it very useful to reset maximum or minimum values. Try, for example, setting the minimum on the y-axis to 0 and the maximum to 100.
+
 
+
There is much more that you can do with Excel, including copying the figure you have just created and incorporating it into a word processing file. Experiment.
+
 
+
Exit from Excel (no need to save the file unless you want to), and click on Continue to move from the Scatterplot back to the Select Variables to Analyze Across Countries window (reproduced near the top of this lesson).
+
 
+
Although the introduction of a control variable works helps in looking at relationships involving up to three variables (dependent, independent, and control), it is possible to examine relationships involving as many as five independent variables (see again the IFs window shown earlier on analyzing variables across countries). The key to this is selection of the Statistics option that you can see at the bottom of that window (next to the Plot button). After selecting a dependent variable and as many as 5 independent variables, touch the Statistics button. That will display the set of transformation options.
+
 
+
http://www.du.edu/ifs/help/use/data/crosssectional/crosssectional.html image
+
 
+
The transformation options allow you to apply transformations to your dependent and/or independent variables, just as you saw earlier in Excel that a logarithmic transformation of GDP per capita helped explain life expectancy better. Choose here again the logarithmic transformation of GDP per capita and then select the button Compute and Show Statistics. That will produce the following window showing statistics about the relationships among these three variables. These statistics will not be explained here – please see a standard statistics textbook.
+
 
+
http://www.du.edu/ifs/help/use/data/crosssectional/crosssectional.html image
+
 
+
'''Computed Functions.'''&nbsp;Click on the sub-option for computed functions to obtain the following window.
+
 
+
http://www.du.edu/ifs/help/use/data/crosssectional/crosssectional.html image
+
 
+
Many two-variable relationships have been computed within IFs using the cross-sectional analysis capabilities described earlier. For instance, relationships have often been computed for the same two variables at different points in time to gain insight about how the cross-sectional relationships have changed over time.
+
 
+
Click the Extend List box in the window above to see the list of computed functions more clearly. Then scroll down until you see GDP/Capita (PPP) Versus Life Expectancy (1962). Because of data availability this was the earliest relationship computed between the two variables. Notice that other relationships have been computed at intervals. Holding the control (Ctrl) button down click on the relationships between those two variables for 1962, 1980, and 1995. All three should be high-lighted and therefore selected. Now click on the Draw Graph option to create a graph that contains all three.
+
 
+
Note that at low levels of GDP per capita at PPP (in constant dollars over time) the life expectancy has generally gone up. In other words, countries have succeeded in raising life expectancy even at low levels of GDP per capita. Why? It could be some combination of improved medical technology and of improved health policy. In any case, the shift in the relationship over time suggests the value of multi-variate analysis of the relationship, not just bi-variate.
+
 
+
Another feature of this section of IFs is the&nbsp;[http://www.du.edu/ifs/help/use/repeated/computations/computations.html Computations]&nbsp;button.
+
 
+
== Cross-Sectional Scattergram ==
+
 
+
This section describes how to use the various features on the Scattergram, which appears after clicking the Plot button in the Analyze Across Countries (Cross-Sectional) option. The descriptions begin with the features found in the upper left-hand corner of the screen and ends with the features on the right of the screen.
+
 
+
http://www.du.edu/ifs/help/use/data/crosssectional/scattergram.html image
+
 
+
Clicking on Continue returns the user to the previous menu, while clicking on Save provides options to save the data as a graph or as values.
+
 
+
The&nbsp;'''Excel'''&nbsp;option will export this data to an Excel spreadsheet.
+
 
+
The&nbsp;'''Display Labels'''&nbsp;feature allows the user to change the labels on the graph. The user can choose to display the full names of the countries or the abbreviations of the names. The user can change the labels of the x- and y-axes, the title, and the subtitle of the graph. The user can also choose to display certain data points and labels and not others by clicking on the Select Points and Labels option. After clicking on Select Points and Labels, a dialogue box appears with a list of all data points and corresponding labels. Next, after choosing which data and labels to display, click Save. The graph should now only display the selected labels and data points.
+
 
+
The&nbsp;'''Display Options'''&nbsp;feature allows the user to access various third dimension display options. One of these options is the Bubble Display feature, which displays the ’ of a given countrypopulation or Gross Domestic Product and facilitates comparison with other countries. This feature can be used to display either the dependent or one of the independent variables. Additionally, the user can change the geographic units that are displayed from Countries to Groups or Geographic Lists. Notice, however, that even when countries are displayed, they are colored according to regions. Click on Define Colors for G-Lists for a key to the various color-groupings. To change the color-grouping method, select a new region from the drop-down menu in the Define Colors option. Then click on Initialize Colors. The colors should now change according to the new grouping method.
+
 
+
The&nbsp;'''Display Dynamic'''&nbsp;option allows the user to create a movie of the interaction of these variables over time. There are two types of movies that the user can view: the first type traces the data points or the dots’ over time, while the second type simply shows the location of the various data points at a given point in time. To view a movie, first click on Display Dynamic and then select the Tracing Mode option. To label only certain or all years, click on Label Years. Click again on Display Dynamic, and then Show Movie, either with tracing on or off. Additionally, the user can display change through time by clicking on the Previous and Next buttons at the bottom of this display. Check Save Movie at the bottom of the screen to review the movie in the future. The movie will be saved under Stored Map Movie, which is located under Specialized Display.
+
 
+
The&nbsp;'''Relationships'''&nbsp;option allows the user to fit a line to the graph. To remove the line, click on No Regression.
+

Latest revision as of 20:17, 27 July 2017

World Map: Historic Data

The World Map is located under the Data Analysis option on the Main Menu of IFs.

World Map example

The main function of the World Map is to display data in a visually dramatic format. Instead of representing country-related forecasts as lines in a graph, or columns in a table, the user is able to view the forecasts as shades of color imposed on a map of the world.

The main features of this option are located at the bottom of the screen which allows users to change the data set that is being displayed. Be sure to click the Refresh Map button after making any changes.

  • Use Continue button to exit and return to the IFs Main Menu.
  • Data availability statistics and source information of each data series are available from the Data Information button.
  • The default data table that is available is TimeSeries, a broad group that contains most longitudinal data series in IFs. By clicking the Data Table drop-down menu, users can select from different data sets. These sets then correspond with a larger list of variables displayed in the Data Field list box.
  • Below the Data Table option is the Dimension of Variable which allows users to change the year being displayed. Options include individual years, earliest year available for each country and most recent year available for each country.
  • Boxes labeled Display Type and Number of Categories allow users to customize the map display. The default display type is set to Equal Interval, which divides the total range of values into the number of categories specified. The alternative categorization, Equal Count, displays an equal amount of responses in each category. For any data display option, users can create up to 7 distinct categories for display.
  • Another input box, Color, determines the shading color in which data is displayed.
  • Finally, by selecting the Projection option at the very bottom of the page, users are able to toggle between multiple global cartographic projections.

Some final features of the map are located inside the map display.

  • The map legend is located to the left of the map. It displays the data points that delineate the different display categories.
  • To move around the map, use the arrow buttons located to the right of the map; to reset the view, click the button at the center of the arrow buttons; zoom in or out using the plus/minus bar.
  • Hovering over each country in the map displays the associated value of the data set, and left clicking on the map allows users to view either a detailed map of the country or a table of the data values by year.

Analyze Across Countries (Cross-Sectional Analysis)

From the Main Menu, an option under Data Analysis provides users with the ability to examine the relationship between historical variables. From the Data Analysis menu, select Analyze Across Countries (Cross-Sectional) to reach the screen displayed below.

Variables across countries window

To do a cross-sectional analysis, specify a dependent variable (a variable you want to understand the causes of) and one or more independent variables (the possible causes of change in the dependent variable). By convention, the dependent variable appears on the y-axis and the independent variable appears on the x-axis on a graph. To select a dependent variable first, scroll through the alphabetized variable list and left-click on any variable in which a user might be interested. Click Select to choose that variable. Users can also type a variable name in the simple text search box directly above the list, which filters series by prefix. Alternatively, there is a more sophisticated search feature available through the Search button at the bottom of the screen. From the search screen, a search on "female", for example, would find all results that contain the word, anywhere in the variable name or description. After selecting a dependent variable, either from the variable list or from the search form, a drop-down list will appear, specifying the year to be displayed. A cross-sectional display includes data for a single year, earliest available or most recent available for each country. The process described here is also used to assign independent variables. To reassign any variable, select the corresponding option button and simply load the new series.

As a demonstration, plot a cross-sectional relationship between female education and reproduction rates. For the dependent variable choose total fertility rates, by selecting TFR from the list of variables. Left-click the variable and two options will appear: Select and Data Information. Click Select to choose that variable. The latter provides detailed information about the scope, availability and source of the data set. The ensuing drop-down list determines the year for which data will be displayed. For this example, choose 2010. Notice that TFR(2010) is then shown as the Dependent Variable. Next, choose an Independent Variable, the level of adult female education, by selecting EdYearsAge15Female for the same year. Click the Plot button to generate the graph displayed below.

Scattergram of raw country data example

To learn how to use various features on the scatter plot, click here.

Cross-Sectional Scattergram

This section describes how to use the various features on the Scattergram, which appears after clicking the Plot button in the Analyze Across Countries (Cross-Sectional) option.

Scattergram of raw country data example

The Relationships option allows users to fit a line to the graph. To impose a line of relationship on the scatter plot as displayed above, choose Relationships from the toolbar, and the option Linear. Notice the r-squared value (a measure of the goodness of fit between the independent and dependent variable) and linear equation displayed at the top of the screen. Additional relationship types include: logarithmic, exponential, power, polynomial and logistic. To remove the line, click on No Regression.

The Display Labels feature allows users to change the labels on the graph. The default selection is No Labels. Although the country names are not shown on the graph with this setting active, users are able to see them by hovering over the data points on the plot area. By switching options under the Display Labels menu, countries can be represented by the full names or the abbreviations of the names instead. Users can also fully customize the labels of the x- and y-axes, the title, and the subtitle of the graph.

Alongside Display Labels, the Specify Geography menu option allows the user to display data by country, group, or geographic list.

Clicking on Continue returns the user to the previous page.

Analyze Across Time (Longitudinal Analysis)

Whereas cross-sectional analysis (Analyze Across Countries) concerns historical relationships between different variables, the primary purpose of the longitudinal analysis (Analyze Across Time) is to trace historical trends of individual data series. This option can be accessed from the Data Analysis option on the Main Menu of IFs.

Variables across time window

To start with, pick any data series to load a dependent variable. Time is set as an independent variable by default. Users can select a variable by scrolling through the dependent variable list box, typing a variable name in the text box or with the more sophisticated, separate Search option, located at the bottom of the screen. Click Select Countries to specify the data that will appear in graphs. (Note that Use Groups , the button beneath Select Countries, is a toggle, switching from country to group selection or vice-versa; e.g. if groups are selected the toggle will read "Use Countries"). From the country selection screen, pick a country, a group or a decomposed group. Buttons on the right side of the country selection screen also allow users to switch aggregation to groups or decomposed groups. Continue back to the original screen for analysis over time and Plot the result.

To demonstrate, pick a variable such as AidDon%GNI (aid donations as a percent of Gross National Income, the successor concept to Gross National Product). Click the Select Countries button and choose United States as a donor country. To generate the graph displayed below, click Plot. You will see the trend of U.S. aid commitment over time.

Graph of Longitudinal Data Example

On the Graphical Display screen as shown above, options on the top menu bar allow users to go back to home screen, alter the scale, display a table, format titles and explore future trends. Under the Trend menu, users can impose extrapolations—linear, polynomial, logarithmic, exponential and s-curve—onto the chart. The sub-option, Extrapolation Setup, features customization options for polynomial, exponential and s-curve trend lines.

An additional feature of the longitudinal analysis is the ability to look more closely at the data set in which a user might be interested. Left-click on any variable and a small window with two options will appear: Select and Data Information. Click Data Information in order to access detailed information as to the scope, availability and properties of the data set in question.

Some other features of this section of IFs include the Computations button and the Table button.

Other than the default setting, treating time as an independent variable, which allows users to look at a historical trend of one variable over time, two other types of analysis are also available in the model: treatment of time As Points and As Time. When the Treatment of Time check box labeled As Points is active, the data points become years and users can look at the relationship between two variables as x- and y-axes. The third option, treatment of time as time, facilitates side-by-side comparison of longitudinal trends with two different variables on the same display (the x-axis is time, in this case).

Exploration of maps, cross-sectional and longitudinal relationships, and relationships computed for IFs can give you much information about the apparent relationships among a wide range of global development indicators. Even if you never used IFs for forecasting, this data analysis capacity could significantly enhance your understanding of the world.

Show Computed Functions

To access Show Computed Functions, click on Data Analysis on the Main Menu of IFs.

By clicking on Show Computed Functions, you will be presented with a menu that looks similar to the one pictured below:

Show Computed Functions menu

The box that is located in the middle of your screen holds hundreds of functions. If you click on the Extend List option right above this box, it will display an expanded list of functions.

In the Filter Functions box on the top of your page, type in any search terms for functions that you would like to see displayed and hit the Enter key. This will bring up all the functions/relationships that have your term(s) in their names.

After you have selected a function, choose the Graph option from the top of the menu and click on Draw. This will then display the function at the bottom of the screen along with the specific function and an R-Squared value.

Select one or more of functions if you would like to see multiple relationships on one screen. In order to select more than one function, highlight the first function, hold the Ctrl key and select a second function.

Identify Group or Country/Region Members

To identify regions or group members used in IFs, click on Data Analysis on the Main Menu of IFs, and then select Identify Group or Country/Region Members.
Identify Group or Country/Region Members window

This feature of IFs allows users to identify which countries/regions are represented in certain groups. By clicking on any of the groups in the left-hand list, the members of that group will appear on the right hand side.

By clicking on Using Regions, the list of groups on the left hand side will become a list of countries/regions.