Copyright © 2025 Terrance Williams
EWI Tracker is an application developed to aid users in tracking and visualizing their workload(s).
Communicating one's workload is a difficult task due to the large number of variables and external factors influencing it. Typically, companies attempt to estimate workload through performance metrics—output is typically the sole factor informing the perception of an employee's workload. However, non-performance-based metrics contribute to one's workload as well. For example, the number of emails, phone calls, and meetings one participates in can take up a substantial amount of one's workday, even if no tangible performance metric results from the effort. Another factor in one's workload, though rarely discussed, is how much time one spends on tasks outside of his or her official role. It is no uncommon for employees to help peers, juniors, and even superiors with their work, especially if the employee in question is a high performer. This volume of work may not be reflected in official performance metrics (ex. helping a peer revise a presentation, troubleshooting an everyday task, or advising during urgent, department-related matters).
Another factor affecting the perception of workload is one's emotional disposition. While this is sometimes acknowledged in passing (ex. "company morale"), it is difficult to establish a standard method of evaluation. This difficulty is exacerbated by:
- Varying levels of emotional intelligence/empathy among leaders.
- Trepidation on part of the employee due to fear of reprisal (direct or indirect).
However, it is important to explicitly acknowledge this component. Someone who is emotionally fulfilled/invested in their job would likely perceive a high technical workload in a more positive light when compared to a more dissatisfied or outright unhappy state. In an especially positive state, the employee may even enjoy the challenge associated with such a load. On the other hand, for an employee who is in a less fulfilled state, it is possible that even a normal workload would be perceived as too much.
With the above in mind, this application attempts to provide a way for employees to track both aspects of workload perception. Though it's difficult to objectively quantify the concept of workload, this program uses a two-part index that the author calls the Employee Workload Index (EWI) to estimate workload intensity. The visualization produced by the program can then be used to give an imprecise-but-useful indication of how a given employee likely perceives their current workload.
EWI is a simple model the author created to aid in workload visualization. The idea is to have two aspects of workload contribution—both a vocation-based technical aspect and an emotion-based personal aspect—and use them to visualize the given employee's workload. It is important to note that this application does not provide a mathematically-proven heuristic for evaluating workload. Instead, it uses information provided by the user to provide a helpful approximation.
The technical aspect of EWI is the more objective of the two. When people evaluate someone's workload, they do so by implicitly comparing the workload to some ground truth or "normal" workload. In other words, technical workload is viewed in this application as a relative concept. The program allows the user to define a series of metrics associated with a job and specify an estimated global average for each metric. Then, the user tracks their values for these metrics over time. Upon request, the program then compares the user's local averages with the global averages.
As an example, let's consider a role in which daily phone call duration contributes to technical workload. Say the global average for the role is 30 minutes (0.5 hr) per day. If Jane Doe averages 2 hours of calls per day for a given calendar period, the program will report that Janes has 4x the average workload for this given metric. Similarly, if Jane averages 0 minutes, the program will report the workload for this metric as 0. Since the author is not prescient enough to know what metrics a given industry may want to define, negative values are also allowed. However, the program was written with non-negative numbers in mind.
The personal aspect is simpler. The user is asked a number of pre-determined questions that aim to assess their emotional satisfaction with their current role. These questions evaluate confidence in leadership, investment in company/department mission, and more. Each question requires a response ranging from 1 to 5 (inclusive). The responses are averaged and normalized to produce a single value ranging from -1 to 1 (incl.). Whether the calculated average falls on the positive or negative end of 0 determines how "satisfied" the user is. This value can then be used to help interpret how the user may perceive their workload found in the technical portion. Also, unlike the technical aspect in which indexes are defined on a per-metric basis, the personal index is one number. The author made this decision in order to help maintain employee privacy. In other words, even if the personal aspect index is negative, a given employer or superior wouldn't be able to know which questions were answered negatively. The personal index is also intended to encourage leaders to communicate with those they lead to help identify unmet needs or the aspects that work well for that employee.
Whether using this application on Windows or Linux, the gnuplot software is
required to generate the EWI plots.
For Windows users, the gnuplot software distributors provide a simple
installer accessed here. Download
the latest version to keep setup simple. When running the installer, keep the
default options until reaching the "Select Additional Tasks" page. On this
page, scroll to the bottom and make sure the Add application directory to your PATH environment variable option is checked (see the image below). It is
important to do so in order for this program to communicate with gnuplot. If
you install gnuplot without checking this option, simply uninstall it and run
through the installation process again.
Linux users can either find gnuplot via their package manager or build from
source.
On Fedora Linux:
$ sudo dnf install gnuplotOn Ubuntu/Debian:
$ sudo apt install gnuplot # This should work, but is untested by the authorEnsure the version is at least 5.2.6.
Finally, Linux users must have access to the Qt5 binaries to run this application. At the time of writing, the author is unfamiliar with distribution for Linux (and, admittedly, is also inexperienced for Windows), so it may be easier to build this application from source.
After that, run the ewiTracker executable to get started.
The EWI Tracker navigation bar shows all of the options available to the
user:
Home: The Home PageProfile: Load the profiles necessary to use the appActions: Record entries and/or produce visualizationsHelp: A "Quick Help" menu to get the user up and runningAbout: Information about this application.Exit: Exit the application, saving any updated changes in the process.
| Nav Bar (Action Button Disabled) | Nav Bar (Action Button Enabled) |
|---|---|
 |
 |
The Profile menu houses options for the app to retrieve information necessary
for its operation. The user must provide both a user profile and a job profile
for the app's actions to become enabled. The user profile consists of basic
information: a user ID and a display name. The ID is intended to be a unique
identifier (ex. your employee ID). The display name can be any name the user
chooses, but be aware that this name appears on the visualization image.
The job profile consists of the identifier (job code + human-readable title) and a list of questions and metrics for tracking. More on the job profile will be discussed in a later section.
In order to create a profile for use, click the Create button located under
the "User Profile" heading. A window will appear requesting information from
the intended user. Once both fields are occupied with text, the "Ok" button
appears. Pressing this button will create the user.
Note: In order to save your newly-created profile, ensure you shut down the
program via the navigation bar's Exit button once you're done with the
session. If this is not done, the program will not be able to locate the user
on the next run. In general, make it a habit to use the Exit button.
To load a previously-created user, click the Load... button under the "User
Profile" heading and input the user's ID.
To load a job profile, select the Select Job... button under the "Job
Profile" heading. This will open a file selection dialog, allowing the user to
select a job profile that's been defined. The program expects profiles to be
stored in its .jobs directory, but job files may be loaded from anywhere.
A job file is simply a text file formatted in a specific way. The first line defines the job's identifier—its formal code and human-readable name. The format is:
<job code>: <human-readable name>
without the angle brackets. Spaces are allowed in the human-readable name. By
"human-readable", I mean the job's title. For example, one role in a restaurant
may have a job code 5555 but the "human-readable" name would be "Head Chef".
A blank line should follow the identifier line.
The rest of the document defines the metrics to track for the role. One metric is defined per line. A given line is formatted as:
<Metric question> | <estimated average>
again, no angled brackets. The | character is typed by SHIFT + \ on a
typical QWERTY keyboard. Each metric has a question that is presented to the
user when making an entry (ex. "How many calls were made?"). It is good
practice to include the expected units in parentheses (see the example below).
The estimated average is the value that will be used to compare the user's
tracked data for that metric. The program expects a number for this value. -
signs are allowed. Decimal values are also allowed.
To summarize, a job profile consists of the following:
- an identifier
- blank space
- Metric question-average pairs.
Here is an example of a fully-formed job profile:
0260: NETC EEO Informal Process Counselor
Time spent on communication (hr)? | 3
Active cases? | 2
Informal contacts? | 2
This results in the following entry survey:
Finally, job profiles can be loaded from anywhere, but it may be easier to
store them with the program itself. Save the file to the <EWITracker location>/.jobs folder. If this folder is not present, try running the
application to ensure the program creates the folder. If this folder is still
not visible, ensure View Hidden Items is enabled on whatever file explorer
you are using. If all options fail, manually create the folder in the
application's directory.
Caution: Unfortunately, once data for a job profile is recorded, the job profile cannot be changed. This is due to inexperience on the author's part as it was unclear how to handle the case where the number/order of questions is changed. This will hopefully be solved in a future iteration.
A user can switch profiles by running through the desired process again. Simply load another job and/or user. If the user is changed by either loading or creating another profile, the previous profile is saved before the new profile is loaded into the program.
The Action menu houses the operations relevant to recording and reporting
metrics. The first heading on the page allows the user to take surveys. To
record technical metrics, take the technical survey. For personal metrics, take
the personal survey. The program will then open a form that the user will fill
out. Technical forms will present the questions defined in the job profile.
Each field only allows numeric input with the exception of "Additional Notes",
a field that allows the user to record any additional information that may be
relevant to that entry.
Note: In the future, the author plans to implement an action that allows
the user to an entry from a specified date. At the time of initial release, the
user can "cheat" by viewing the desired entries in their file located in the
.usr directory.
Taking a survey is as easy as clicking the desired survey button. The form is programmed such that it will only allow submission once all metric fields are populated. The "Additional Notes" field can be empty. It is important to note that the program expects and enforces chronologically-ordered entries. In other words, each technical entry must have a date that is later than the previous entry. If the user intends to submit multiple entries (say if they hadn't been able to access their computer for some time), the entries should be made from earliest to latest date. The same rule applies to personal entries. However, technical and personal entries are independent of one another; the program does not enforce chronological ordering between the two.
The frequency with which surveys are taken depends on both the user and the job profile definition. Some metrics may be written to be recorded daily; others may have a longer timescale. Personal surveys are intended to be taken weekly, but the program does not enforce this. Frequency is at user discretion.
To view the EWI report, click the View Workload... button under the "Metrics"
heading. The program will then ask for the desired date range for the report.
It is an inclusive range, so if both the From and To fields have the same
date, the program will generate a report for that day if an entry exists for
it. Depending on your machine, the plot may take some time to display.
The following is an example report based on contrived metric values for the job defined earlier in this document:
Let's discuss how to interpret the plot. If the user only recorded technical data for a given date range, only the technical plot will be displayed. If the user had entries for both technical and personal data within the date range, the program will output a two-part plot. The left portion will have the technical data, and the right will have the personal data. If there is no technical data at all for that date range, no plot is generated.
The technical plot consists of a baseline at y=1 and a series of scatter
points. The baseline represents the global average for each metric; since each
metric can have varying value scales, the program normalizes each metric such
that the technical index is 1.
A given scatter point (x, y) represents (Question num, Technical Index),
where question number 1 corresponds to the first question on the job profile.
The Technical Index (Idx) is the workload approximation. A y value of 4 for
example means the user has 4x the average expected workload for that metric.
Knowing this, we can see visually how much the user's workload deviates from
the norm at a glance for each metric.
The right side of the two-part plot displays the Personal Index. This plot is simple to interpret; where does the non-black point fall on the vertical line? If it is above the black point, the user's responses indicated a positive disposition with intensity increasing the closer the index is to 1. If below the black point, the responses indicated a negative disposition. Finally, if the non-black point completely overlaps the black point (Idx=0), the responses indicated that the user is exactly neutral.
The user can save the generated visualization to their file system via the
Export button.
At the time of writing, there is no functional export feature for this app; the user must do it manually.
In short:
- Locate where the program executable is located.
- In your file explorer, ensure
View Hidden Itemsis enabled. - Copy the
.usrfolder to backup users and the.jobsfolder to save jobs.
- cpperrors (0.1.0)
- Eigen (3.4.90)
- Matplot++ (1.2.1)
- gnuplot (at least 5.2.6)
- Qt (5.15.2)
- CMake (3.30.5)
- vcpkg (2024-11-12-eb492805e92a2c14a230f5c3deb3e89f6771c321)
- Clang (19.1.5)
- gcc (14.2.1)
- Ninja (1.12.1)
- Visual Studio 2022 (Community Edition)