DSC 330 Week 3 Building Visuals and Dashboards Example

From Sketch to Screen: Building a Hotel Company's Dashboards in a Business Intelligence Tool

[Student Name]

University of Phoenix

DSC/330: Data Communication and Visualization for Business

Week 3 Assignment

[Instructor Name]

[Date]

Music Row Hospitality, its hotels, data and figures are composites written for a model paper.

Weeks 1 and 2 gathered requirements and chose charts for Music Row's dashboard, covering the nine properties it manages. The first release has two views: a weekly portfolio screen for the operating chief and a phone-sized daily screen for each general manager. This paper describes how both views were assembled in the analytics software the company already licenses.

Dashboards as a Genre

Sarikaya et al. (2019) analyzed a large collection of real dashboards and found that they serve varied purposes, from strategic decision making to operational monitoring to communication, and that dashboards differ in interactivity, update frequency and audience. They argued that designers should be explicit about which kind of dashboard they are building. The portfolio view is strategic and weekly; the general manager's view is operational and daily. Their designs therefore differ.

Connecting and Preparing Data

Each hotel's property management system exports a nightly file of reservations, rooms sold, rates and revenue. The build loads these files into the tool's data model, along with weekly guest survey results, payroll hours by department and the annual budget by hotel and day. Two problems surfaced. Three hotels count complimentary rooms as sold and others do not; the model excludes complimentary rooms everywhere, a rule agreed with the revenue manager. One hotel's export uses different department codes; a mapping table standardizes them.

Defining Measures

Precise definitions keep numbers consistent with other reports. Revenue per available room equals room revenue divided by rooms available, excluding rooms out of service. Labor cost per occupied room equals hourly payroll cost for rooms-related departments divided by rooms sold. Satisfaction equals the mean overall score on a 10-point scale, shown only when a hotel has at least 30 responses in the period, since smaller samples swing widely. Each definition appears in an information panel on the dashboard.

The Portfolio Dashboard

Few (2006) recommended placing the most important information where the eye goes first, grouping related items and removing anything that does not inform. The portfolio dashboard follows that advice. At top left, a sorted bar chart shows revenue per available room for the week by hotel, with budget markers. To its right, a small table shows each hotel's variance from budget and last year. Below, a dot plot compares satisfaction this week and the same week last year, and a bar chart shows labor cost per occupied room against target. A single filter selects the week; hotels below target on two or more measures are marked in the accent color.

The General Manager's Dashboard

The daily view is a single column sized for a phone. At top, yesterday's three headline figures, rooms filled, average rate and revenue compared with plan, each with an arrow showing direction from the prior week. Below, a line of the last 30 days of occupancy with a budget band. Below that, bars count the week's guest complaints for each department. Tapping a department shows the comments. The view opens to the manager's own hotel and cannot show others, as owners required.

The general manager's whole dashboard fits on a phone screen before the morning meeting, which was the only requirement that mattered to them.

Interaction

Card et al. (1999) describe interaction techniques such as filtering, details on demand and linked views as ways to let users explore data without overwhelming them with everything at once. The dashboards use interaction sparingly: one week filter on the portfolio view, drill-down from a hotel bar to that hotel's daily detail and comments on demand in the daily view. More filters were tempting but would make the dashboard harder to use quickly.

Performance

The first version took 14 seconds to load on a phone because it pulled a year of daily data. Limiting the daily view to 30 days and precomputing weekly totals reduced load time to under three seconds.

Security and Access

Owners see only their own hotels, so the tool's row-level security filters every view by the signed-in user's hotels. General managers see their own hotel only. The chief operating officer and revenue manager see all nine. Access was tested by signing in as each role before release, which caught one view that briefly showed all hotels to an owner's account.

Usability Testing

Five users, the chief operating officer, two general managers, the revenue manager and an owner's representative, were each given five questions to answer with the dashboards, such as "Which hotel had the biggest drop in satisfaction from last year?" and "Was yesterday's revenue above budget?" Observers noted hesitations and errors. Four findings led to changes. The label "RevPAR var." confused the owner's representative; labels now use plain words. A general manager tapped the wrong department because bars were too thin on the phone; bars were widened. The chief operating officer wanted last year's figure visible without hovering; it was added to the table. The revenue manager asked for competitor rates, which are planned for the next release.

Documentation and Hand-Off

A dashboard that only its builder understands will decay. The build includes a short guide covering data sources, refresh schedule, measure definitions, the mapping table for department codes and how to add a new hotel. The analyst who maintains the monthly report has been trained to update the dashboards, and a change log records every modification so that numbers can always be traced to a version of the rules.

Tool Limits

The tool handles the data volume easily but offers limited control over mobile layouts, so the phone view required a separate page rather than an automatic resize. Its color palette defaults to many hues; a custom theme restricts colors to gray and one accent.

Conclusion

Building the dashboards meant connecting and cleaning data from nine hotels, defining measures precisely, laying out views by importance, adding only the interaction users need and testing with real questions. Research on dashboard genres and design guided choices, and testing produced four changes before release. Week 4 will validate the data behind the dashboards.

References

Card, S. K., Mackinlay, J. D., & Shneiderman, B. (Eds.). (1999). Readings in information visualization: Using vision to think. Morgan Kaufmann.

Few, S. (2006). Information dashboard design: The effective visual communication of data. O'Reilly Media.

Sarikaya, A., Correll, M., Bartram, L., Tory, M., & Fisher, D. (2019). What do we talk about when we talk about dashboards? IEEE Transactions on Visualization and Computer Graphics, 25(1), 682-692. https://doi.org/10.1109/TVCG.2018.2864903