CEMEX Manufacturing Dashboards
Data Visualization
User Research
From 10+ Excel tables to KPIs ready for every meeting
CEMEX’s regional leaders reviewed energy, CO2 and production in Excel reports before important meetings. As the only UX designer, I turned those reports into Power BI dashboards for 9 plants in Spain and Germany.
Role
UX Designer
Timeline
Late 2024 – Mid 2025
Team
1 UX designer, 2 Power BI developers, 1 PM, 1 analyst
Tools
Figma
Claude
Power BI
Problem
The numbers that mattered were buried in rows
Regional leaders answer for these numbers, but they lived in Excel reports spread across 10+ tables from different sources.
Before important meetings, leaders searched rows that all looked alike for their KPIs and rebuilt them in slides. When something went wrong, they went back to the same tables to find the cause.
In a first round with the Mexico team, keeping this data up to date by hand took each person about 2 hours.
One of the Excel reports used before the dashboards.
Process
From interviews to validated views
I was the only UX designer, with 2 Power BI developers, a project manager and an analyst.
My approach, in three steps:
Interview
A first round of questions with the Mexico team, then interviews with 3 regional leaders in Europe.
Synthesize
Interview notes analyzed with Claude and checked against the originals, to find behavior patterns.
Design and validate
30+ views, validated with 6 plant users in Spain and Germany.
Interview notes
What they said, and what they did
In every interview, we wrote down what leaders said about the reports and what they actually did with them. Those notes became the patterns behind the findings.
Notes taken over the stoppages report during the interviews, in Spanish.
AI in the process
Where AI sped up the work
I used Claude as a research partner and checked every output against the source.
Patterns from interviews
Claude grouped the interview notes into behavior patterns. I checked each one against the original notes before it shaped a decision.
Faster iterations
The same patterns guided each round of design changes.
Overall findings
What the reports were really for
Three leaders, similar roles, different interests in the same data.
Meetings drive the work
What we saw
Before key meetings, leaders searched look-alike rows for their KPIs and rebuilt them in slides.
Design response
Main views show the KPIs, ready to present.
Three numbers lead
What we saw
Energy consumption, tons of CO2 emitted and tons of cement produced.
Design response
Those KPIs open every main view.
Detail still matters
What we saw
They compared exact values for each mill and plant at the same time.
Design response
Some tables stayed, redesigned for reading.
Two regions, two priorities
What we saw
Germany focused on consumption; Spain, a bit more on production.
Design response
Separate views for each region, built around its own priorities.
Design direction
Lead with the KPIs, keep the detail within reach
Each main view starts with the KPIs leaders answer for: energy consumption, CO2 emitted and cement produced. The detailed tables stayed, redesigned for the questions a chart can’t answer.
Explorations
Sketching before building
Before opening Power BI, I sketched ways to show the same data: KPI cards, trends, tables and distributions. The notes already pointed to the final decisions: less load, and equipment detail over whole-phase totals.
Whiteboard and paper sketches, before opening Power BI.
Key decisions
Charts first, but some tables stayed.
Why
Leaders needed exact values for each mill and plant side by side, and no chart could replace that.
Principle
Cognitive Load
Evidence
We kept some tables and redesigned them: clearer row order, larger type and statuses.
KPIs over totals.
Why
Leaders answer for their own KPIs. The total produced by all plants together didn’t help them decide.
Principle
Selective Attention
Evidence
After validation, main views lead with KPIs instead of bar and pie distributions.
UI aesthetics
Charts that read at a glance
The old charts competed with the data. The new ones let it lead.
Before
Before: an Excel chart from the old reports.
After
After: a chart from the dashboards.
What changed
1
Saturated default colors → a calm palette where color marks meaning.
2
Crowded bars without values → values on the bars and a clear legend.
3
A cramped plot → space around each chart.
Principle
Aesthetic-Usability Effect
Selective Attention
A clean chart is read faster and trusted more, and these leaders present the numbers to others.
Before and after
From Excel rows to one stoppages view
Stoppages lived in Excel as rows of numbers, times and causes. Now the view opens with today, period and year-to-date totals, then equipment status and trends, with the detailed table still in reach.
Before
Before: the stoppages report in Excel.
After
After: the plant stoppages overview for Germany.
Final result
Two fully navigable dashboards, each built around its region’s priorities
Spain and Germany each got their own views, shaped by what each region tracks most: production in Spain, consumption in Germany. Both follow the same structure, KPIs first and the detail within reach.
A selection of views from the Spain dashboards.
A selection of views from the Germany dashboards.
After launch
What the dashboards were built to change
Expected outcomes, backed by what users told us. Time saved was not measured.
Less time getting ready for meetings.
Why
KPIs are ready to present, so leaders no longer rebuild them in slides.
Principle
Interaction Cost
Evidence
Plant users said their work became much faster. Before, the Mexico team estimated about 2 hours per person to update this data by hand.
Faster answers when something goes wrong.
Why
Status, trends and the detailed table sit in one view, so tracing a cause takes fewer jumps between files.
Principle
Information Scent
Evidence
Users said decisions based on the dashboard values became more effective.
Impact
Validated in Spain and Germany, then rolled out
Measured
30+
dashboard views for critical KPIs
9
plants in Spain and Germany before the rollout
Qualitative
Plant users said their work became much faster, and that decisions based on the dashboard values became more effective.
Used mainly for monthly sustainability reporting, and rolled out to the US, Mexico and the rest of Europe after validation.
Note:
Feedback is qualitative and comes from the 6 validation users.
What I learned
Summary and detail, together
With dense data, not everything fits in a clear visual. Decisions often need both: the summary to see the pattern and the detail to act on it.
Start from the meeting
What people have to present and defend decides what goes first, not the data.
TL;DR
The short version
Problem
Regional leaders searched 10+ Excel tables for their KPIs and rebuilt them in slides before every important meeting.
Research
Interviews with 3 regional leaders showed the reports were for meetings: three KPIs lead, and exact detail still matters.
Design
30+ Power BI views, KPIs first with redesigned tables in reach, validated with 6 plant users in Spain and Germany.
Impact
Dashboards for 9 plants in Spain and Germany, then rolled out to the US, Mexico and the rest of Europe.
After launch
Built to cut meeting prep and speed up finding causes. Users said their work became much faster; time saved wasn’t measured.
