Small confession the first free sample I ever built, I got the main number wrong. I was so focuse...Small confession the first free sample I ever built, I got the main number wrong. I was so focuse...
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Small confession the first free sample I ever built, I got the main number wrong.
I was so focused on making the dashboard look clean that I didn't double check one of the formulas properly. Sent it over proudly, and the client (very politely) pointed out the total didn't match their own records.
Could've been embarrassing. Instead it became the most useful mistake I've made so far. Now every single project goes through the same rule before it leaves my laptop: does this number match a number the business owner already knows to be true? If yes, I trust the rest. If no, I go back and find out why.
A dashboard that looks impressive but is wrong is worse than no dashboard at all. It just makes bad decisions look confident.
Social posts we designed for Coursewell, an LMS for online academies.
We built them like Meta ads: one idea per post, a big headline, and one real piece of the product blown up in the middle, like the grading queue, a cohort's pace or a new order coming in. Ember orange fills only two of the fourteen.
The dashboard itself is available as a UI kit on Figma Community.
NearMiss Transit: spotting small transit delays before they snowball
The problem
• Delay alerts usually fire after a line has already broken down
• Small delays that repeat on different routes can chain together
The idea
• Look for repeating "near-miss" delays instead of waiting for a failure
What it does
• Finds recurring delay patterns on a route
• Links related routes into chains
• Replays a delay timeline: start, peak, recovery
• Workspace dashboard with API docs
Where it stands
• A project demonstration, not client work
• The screens below use historical replay, not live service data