EA-ILES — East African Infrastructure Logistics & Execution System

A road corridor monitoring dashboard, built from a single conversation about a live tender.

Live demo: https://2glj4k.csb.app/ · Case study below

The Problem, In One Sentence

A road contractor running an 18-week paving contract can be quietly falling behind schedule for weeks before it becomes a penalty-triggering event — and by the time a monthly report flags it, the money's already lost.

What This System Actually Does

• Flags a 5.5km schedule variance and 38.3 hours of downtime before they reach a monthly report — the difference between reallocating crews in week 3 versus discovering a problem in week 15
• Turns idle equipment into a visible cost, not a hidden one — engine-running, zero-output hours are surfaced directly, not buried in a utilization chart someone has to interpret
• Converts weather from "bad luck" into a plannable risk — a rainfall/downtime pattern view lets a contractor pre-schedule maintenance into high-risk weeks instead of losing hours reactively

If you're a contractor, this is the difference between finding out you're behind schedule from a spreadsheet a month later, versus seeing it happen on a Tuesday and doing something about it before it costs you a penalty clause.

The Origin

A friend was awarded a road construction tender in Kenya — the Mombasa–Nairobi corridor. He mentioned it in passing; no data, no brief, no ask. This system was built from that single comment: a self-directed exploration of what a monitoring system for a project like his would actually need to do, and a chance to practice the full BI workflow — schema design, data modeling, dashboard delivery — against a real-world scenario instead of a synthetic dataset.

All data shown is illustrative, generated to match a plausible real project shape. The system is built to connect to live site logs.

Problem Framing & Business Constraints

Road construction contracts are usually milestone-based, with penalty clauses tied to cumulative progress by date. Three operational realities shaped the design:

• Schedule risk compounds silently. A contractor can be a few days behind in week 3 and not notice the trend until it's a penalty-triggering event in week 15. The dashboard needed to surface variance against baseline continuously, not just at reporting intervals.
• Equipment cost is a mix of visible and hidden spend. Fuel and rental costs are tracked, but idle time — engine running, no output — is often invisible until someone audits hours manually. This needed its own surfaced signal, not just a bar chart someone has to interpret.