Back to projects
4GitHubLinkedIn

Inventory Health Dashboard for a LATAM Supply Chain Team

Turning a daily Incorta email into a live traffic-light view of safety stock vs actual stock across LATAM.

No AI, no fancy optimizer. Just wiring up the basics so people stop flying blind on inventory.

What this actually is

This is an Inventory Health Dashboard for a regional supply chain team in LATAM.

Every day, they get an export from an upstream analytics system (Incorta) with:

  • on-hand stock by material and country,
  • safety stock levels,
  • some basic metadata.

Instead of:

  • manually opening that file,
  • filtering it a hundred different ways,
  • and guessing where the problems are,

this project turns that daily email into:

  • a Power BI dashboard,

  • refreshed automatically,

  • with a traffic-light status for each material and country:

  • green: stock comfortably above safety stock,

  • yellow: getting close,

  • red: below safety stock, attention needed.

You can slice it by:

  • country,
  • material code,
  • status (green / yellow / red),

and quickly answer the question:
“Where are we exposed right now?”

Why this needed to exist

The starting point was very basic but very real:

  • planners were getting a report in their inbox every morning,
  • opening it in Excel,
  • scrolling around until they spotted something scary,
  • and then reacting.

There was no:

  • consistent definition of “healthy” vs “risky” stock,
  • portfolio-level view by country,
  • quick way to explain to leadership where the real problems were.

It was all:

  • “I think we’re okay in Country A,”
  • “We might be tight in Country B,”
  • and a lot of “let me check and get back to you.”

This dashboard is my attempt to fix that with the least amount of drama:

  • use the report they already have,
  • automate the boring part,
  • give them a single place where they can see what actually matters.

What actually worked

A few things clicked nicely:

1. The traffic-light logic made conversations simpler

Instead of debating raw numbers, people started talking in a shared language:

  • “How many reds do we have in Country X?”
  • “Why is this material stuck in yellow for the third week?”

The exact thresholds are simple, but having them codified in the model turned vague concerns into concrete questions.

2. The automation chain quietly removed a daily chore

The data pipeline is simple but effective:

  • the Incorta report arrives by email,
  • Power Automate detects the attachment and drops it into a SharePoint folder,
  • Power BI is wired to that folder and refreshes on schedule.

Nobody has to:

  • download attachments,
  • rename files,
  • or remember to refresh the dashboard.

It just shows up.

3. The “one-page view” of LATAM changed the discussion

Before this, each country was kind of its own universe.

With the dashboard:

  • you can see all countries side by side,
  • compare how many materials are in red or yellow,
  • and spot patterns, like:
    • a specific material that is chronically understocked everywhere,
    • or a country that always sits on too much inventory.

It turned “how are we doing?” from a feeling into a picture.

Where it is still rough (being honest)

This is not some perfect, global control tower. A few limitations:

  • Still tied to a single upstream report
    If the Incorta layout changes, everything breaks. There is no robust schema control or validation layer yet.

  • Safety stock is taken as a given
    The dashboard treats safety stock as truth.
    It does not question whether those levels are well calibrated, which is probably where a lot of value is hiding.

  • No forward-looking view
    Everything is “as of today.”
    There is no demand forecast overlay, no lead time view, and no estimate of how fast a green could become a red.

  • No formal alerting yet
    The visual is strong, but there is no automated alert like:

    “Send me an email when a red appears in these key SKUs.”

So: it is a big step up from raw Excel, but it is still V1.

What I would do next

If I had another chunk of time on this, I would push it in three directions:

  1. Make the data more robust

    • add a small validation step so if the upstream file changes structure, the pipeline fails loudly instead of silently giving wrong numbers,
    • log ingestion history so you can see which days did or did not refresh correctly.

  2. Question safety stock instead of just visualizing it

    • pull in demand history,
    • recompute candidate safety stock levels,
    • and show where the current settings look too conservative or too risky.

  3. Add alerts and simple workflows

    • let users subscribe to specific materials or countries,
    • send a daily or weekly summary like:

      “3 new reds in Country A, 2 resolved in Country B.”

Technical details (for people who care about the plumbing)

  • Tools

    • Data source: daily Incorta export delivered by email,
    • Orchestration: Power Automate (email to SharePoint),
    • Storage: SharePoint folder as the raw file landing zone,
    • BI layer: Power BI.

  • Data flow

    1. Incorta sends a daily report (typically Excel or CSV).
    2. Power Automate flow:
      • triggers on new email from a specific sender / subject,
      • saves the attachment into a dated folder in SharePoint.
    3. Power BI:
      • connects to that SharePoint folder,
      • uses a Power Query step to:
        • clean column names,
        • select relevant fields (material, plant/country, on-hand quantity, safety stock, etc.),
        • add a calculated “status” column (green / yellow / red).

  • Dashboards and logic

    • Core calculated measures:
      • StockGap = OnHand - SafetyStock,
      • Status =
        • green: StockGap well above threshold,
        • yellow: StockGap close to zero within a band,
        • red: StockGap below zero.
    • Visuals:
      • summary cards (count of greens, yellows, reds),
      • table by country and material with status colors,
      • slicers for:
        • country,
        • material code,
        • status.

  • My role

    • designed the status logic and thresholds,
    • set up the Power Automate flow and SharePoint structure,
    • built the Power BI model and visuals,
    • iterated with the team based on how they actually used it day to day.

I cannot show the real data for obvious reasons, but the architecture and ideas are the important part.
If you are curious about the layout or want a generic version of the PBIX, reach out and I can sketch a sanitized example.