Fathom Tutorial
Fathom™ is a statistics software package that offers a variety of powerful data analysis tools in an easy-to-use format. This tutorial introduces the most
basic features of Fathom™: entering, displaying, sorting, filtering, and manipulating data.
When you enter data into Fathom™, it creates a collection, an object that contains the data. It can then use the data from the collection to produce other
objects, such as a graph, table, or statistical test. These secondary objects display and analyse the data from the collection, but they do not actually contain
the data themselves. If you delete a graph, table, or statistical test, the data still remains in the collection.
Fathom™ considers a collection as a set of cases. Each case in a collection can have a number of attributes. The Case Table feature displays the cases in a
collection in a format similar to a spreadsheet, with a row for each case and a column for each attribute. You can add, modify, and delete cases using a case
table.
We will be using the New Car and Truck 2004 data in fathom to produce appropriate one variable graphs for each of the variables in the data set.
Variable Label Definition Type
Make Manufacturer of the vehicle Qualitative, Nominal
Model Name of the vehicle Qualitative, Nominal
Origin Country of origin for Manufacturer Qualitative, Nominal
Type of Vehicle
Type (Sports Car, Sport Utility Vehicle, Wagon, Minivan, Pickup, Other) Qualitative, Nominal
Drive Drive Train (All-Wheel Drive, Rear-Wheel Drive, Front-Wheel Drive) Qualitative, Nominal
Manufacturers Suggested Retail Price, what the manufacturer thinks the vehicle is worth, including adequate profit
MSRP for the automaker and the dealer (U.S. Dollars) Quantitative, Continuous
Dealer Cost what the dealership pays the manufacturer (U.S. Dollars) Quantitative, Continuous
Engine Size Measured in litres Quantitative, Continuous
Number of Cylinders N/A implies a rotary engine rather than a piston engine with cylinders Quantitative, Discrete
HP Horsepower Quantitative, Continuous
City MPG Fuel Consumption on city streets in miles per gallon Quantitative, Continuous
Hwy MPG Fuel Consumption on the highway in miles per gallon Quantitative, Continuous
Weight Measured in pounds Quantitative, Continuous
Wheelbase Measured in inches Quantitative, Continuous
Length Measured in inches Quantitative, Continuous
Width Measured in inches Quantitative, Continuous
5 Year Resale Value Approximate value of the car after 5 years based on Blue Book Values Quantitative, Continuous
JD Power and Associates Rati ng for Overall Dependability
based on verified owner-reported problems in the first 3 years of new-vehicle ownership that have caused a
Dependability Rating complete breakdown or malfunction of any component, feature, or item. Quantitative, Discrete
JD Power and Associates Rati ng for Overall Performance and Design
Performance and Design Rating Identifies owner delight with the design and performance of their new vehicles Quantitative, Discrete
JD Power and Associates Rati ng for Overall Quality
Initial Quality Rating Highlights production quality and functionality among new vehicles. Quantitative, Discrete
Launch Fathom and edit the preferences changing the font size to largest – this is useful if you plan to copy and paste your fathom work into a
presentation application.
Open the data set with the filename New Cars and Trucks Data 2004.ftm found in the 4studnets folder. The collection appears. The
collection has a generic name. Double Click on the collection name and rename the collection “New Cars and Trucks Data 2004”. Save this file to
your workspace.
Select the collection to make it “active”. Drag the case table icon from the shelf to the work area. The data is contained in the case
table. Scroll down to determine how many cases are there:_____.
The second variable/attribute is called “MakeModel”, double click on the attribute name and change the name to “Model”
From the Collection Menu select Prevent Changing Values in Graphs.
Minimize the table by resizing it, it will appear as an icon when minimized.
TIP: You will need to organize your workspace as you proceed. Group the graphs from each example, in a row for instance. When you are creating a graph the
inspector must be visible. To open the inspector double click on the collection. The variables (aka attributes) are in the case tab of the inspector.
To copy objects from fathom to paste into another application, select the object and press shift ctrl c to copy.
Example 1: Graphing quantitative continuous variables.
With the inspector open, drag the graph icon to the work area.
Drag the MSRP attribute from the case table to the horizontal axis of the graph.
The type of plot can vary for this type of variable. Change the plot from dot plot to histogram and then to box plot using the pulldown menu at the top
right corner of the graph.
Create boxplots for Dealer Cost and for 5 Year Resale Value.
Example 2: Manipulating Data
Create another attribute called Profit. Scroll to the bottom of the attribute list in the inspector. Click on and type Profit.
Calculate values for Profit. Right click on the attribute and select edit formula. Click on the plus sign beside Attributes, Double-click on MSRP then
click on the subtract symbol and double click on Dealer Cost. Click OK.
Graph a boxplot for Profit.
MSRP 5 Year Re saleValue
Create another attribute called Percentage Decrease in Value. Enter the formula 100 . Create a histogram. To
MSRP
determine the average percentage decrease, right click on the graph and select plot value. Type in mean( ).
Create histograms, dotplots or boxplot for each of the following quantitative, continuous variables: Engine Size, HP, city MPG, Hwy MPG, Weight, Wheelbase,
Length, and Width.
Example 3: Graphing quantitative discrete variables
The distributions of a discrete variable are more appropriately displayed with a bar chart rather than a histogram. Click on the Number of Cylinders
attribute and drag the variable to the x axis. Hover the mouse over any category (bar) on the graph and look at the status bar at the bottom left
corner, what do you notice?
Some Quantitative variables may not be discrete (ie have decimal values) but may still be best treated as such due to a repetitiveness in the values of
the variable. For example the three vehicle rating variables included in this set take on values 1-5 but may also include half ratings as well such as 3.5
for instance. Create a histogram for one of the ratings, the notable gaps between each value indicates that this graph is not suitable. Fathom needs to
be tricked into believing that the quantitative variable is a qualitative variable. Click on the one of the ratings attributes then hold down the shift key
and drag the variable to the x axis. Create a bar for each of the vehicle ratings.
Example 4: Graphing Qualitative Variables
If the categories are ordinal they should appear in logical order on the x-axis, if the categories are nominal they will appear in alpha order. Bars can be
rearranged by dragging the categorical name left and right along the graph.
Drag the graph icon to the work area and then drag the Make attribute to the x-axis. Right Click and Sort Bars.
Using the same process, create a bar chart for the Origin attribute. Drag the Origin attribute the x-axis and move it to the y-axis. This swaps the x
and y –axis.
Now create sorted bar chart on the x- axis for the attribute named “Type”, change this graph to a ribbon chart.
Create a bar chart for the attribute called “Drive”. Drag the bars by their categorical names on the xaxis to place the bars in the order –Front, Rear,
All Wheel Drive.
" Domestic"
Create a new Qualitative variable called Domestic vs Foreign. Enter the formula, (including quotes), if(Origin = “USA”) . Make a bar chart
" Foreign"
for this variable.
Example 5: Applying formulas to Bar Charts
When you put a categorical attribute on a graph, you get a bar chart. By default, the height of each bar reflects the number of cases in that category, and the
formula count() appears below the plot area. By editing this formula, you can make the bars reflect some other statistic.
Duplicate the bar chart for the Domestic vs Foreign attribute by right click select duplicate graph.
Double-click the formula count(). The formula editor opens.
Enter the formula you want computed in the graph, for example, type in mean(5 Year Resale Value) using the +attributes to enter the variable
rather than typing it in. Press Enter or click OK to accept the formula and close the formula editor. The bar chart show how the average resale
value varies between Domestic and Foreign vehicles.
If you hover the mouse over any bar in a Fathom bar chart Fathom will read the number of cases and the percentage for that category and display these values
in the bottom left corner (the status bar) of the workspace. Try this using the “Type” bar chart already created. Duplicate this bar chart and follow these
next steps to change the formula so that the percentages are displayed rather than the number of cases.
Double-click the formula count().
With count() highlighted in the formula editor, click the division symbol and then type GrandTotal in the denominator. Select the ratio and multiply
by 100.
Example 6: Using Filters
Duplicate the histogram for 5 Year Resale Value.
Graph only those vehicles with a dependability rating greater than 3.5. With the histogram active/selected, right click and Select Add Filter. Click on
the plus sign beside Attributes, Double-click on the Dependability Rating attribute, select the greater-than button, and type 3.5. Click the Apply
button, and then OK button. Note that the Filter is listed at the bottom of graph. Use the mouse on the y-axis to drag the scale up to achieve a better
scale for this graph.
Duplicate the bar chart for Dependability Rating.
Graph only the sports cars. Select the bar chart then right click and select Add Filter. Click on the plus sign beside Attributes, Double-click on the
Domestic vs. Foreign attribute, select the = button, and type “Foreign” (including quotes). Click OK. Use the mouse on the y-axis to drag the scale up
to achieve a better scale for this graph.
NOTE: Filters may be quantitative or qualitative and can be applied to any type of variable and can be applied in combination.
If all graphs require filtering then the filter can be applied directly to the case table rather than to each graph.
Filters can be edited and removed.
Exploring the Results:
When you select cases in a collection, case table, graph, or summary table, those cases appear as selected everywhere they are visible, providing a simple but
powerful technique for exploring your data.
With the Domestic vs Foreign graph selected, select the portion of the graph corresponding to the “Domestic” response. These cases are now highlighted in
red in all the other graphs allowing insights into how the this variable have been distributed amongst the other variables.
To deselect cases, click inside the graph.
Next steps: describing graphs and calculating summary statisitics.