← Back to dashboard

BrightHaul Logistics dashboard - case study

A live operations dashboard for a fictional regional carrier. Built in about four hours, hosted on AWS for under $1/month, designed to be wired to a real TMS in a couple of days.

Walkthrough — KPIs, live loads, refresh cycle, CSV export

What's actually here

The dashboard reads from a synthetic dataset that mutates every 8 seconds. Six Chart.js visualizations cover the metrics an ops manager at a small or mid-market carrier would actually look at: revenue trend, on-time delivery vs. target, top lanes by revenue, load status mix, fuel cost vs. the DOE regional average, and driver utilization.

The live loads table refreshes on the same cycle and supports CSV export. The "Connect your data" section at the bottom is the bridge - it shows the buyer how the same UI would plug into their actual source.

Why this design

The look is intentional. Most dashboard demos on Upwork are either pulled from a Bootstrap admin template (looks like 2014) or built with a heavy framework like Material that doesn't fit a logistics buyer. This one borrows from the visual language of tools the buyer already pays for: Linear, Mercury, Stripe Atlas. Dark navy panels, a single accent color, monospaced numerals, no gradient hero shots.

The information density is also intentional. Ops people want to see four KPIs above the fold without scrolling, six charts in two rows, and a live-data table they can sort. They are not buying delight. They are buying "I can run my Monday meeting from this screen."

The stack

Wiring it to your data

The synthetic feed is one file: dashboard.js. Replacing it with your real source is a single adapter:

• TMS (McLeod, TMW, Samsara, KeepTruckin): their REST API or a nightly CSV drop into S3, then Lambda transforms it into the dashboard's JSON shape
• Shopify, WooCommerce, Square: for shipper-side dashboards. OAuth + webhook subscriptions, ~6 hours
• Airtable, Google Sheets: the fastest path. Public read API, dashboard polls every 60s. ~3 hours
• Postgres, MySQL, BigQuery: read-only credentials, an API Gateway + Lambda layer in between so the DB never touches the public internet. ~8-12 hours
• Custom REST or GraphQL: one adapter file. ~4-8 hours depending on schema fit

What this would cost as a real project

For a buyer who wants this dashboard, branded for their company, wired to one real data source, with a login and three user accounts, the realistic budget is $1,800 - $3,200. That's roughly 12-20 hours of my time.

The variable is the data source. A clean REST API or Airtable is at the low end. A legacy TMS with a cranky XML feed is at the high end. The dashboard UI itself is the same.

What you would NOT pay for

• A framework license. This is plain HTML/CSS/JS, not Tailwind UI Pro or Refine.
• A dashboard SaaS subscription. You own the code, you host it.
• Per-user fees. Add accounts at the cost of writing the row to your own auth table.
• Vendor lock-in. The whole thing is ~600 lines. You can hand it to any developer.