AgTech

Augmenting data classification

Improving data quality, classification, & uploading

Results

We successfully launched the application in under 3 months, while improving data quality, seed name accuracy, and back-end efficiency.

Improved data quality

Rearchitected the front-end to decrease data quality issues

Faster & more accurate

Increased speed and accuracy of applying seed names to multiple fields at a time

Efficient, orderly back-end calls

Fine-tuned seed identification and old-to-new mapping

Background

San Francsico

When seed demand outpaces supply, farmers need to quickly choose and manually classify new seeds to plant. This means huge data quality and seed recommendation issues for AgTech companies. This impacts 30% of fields in the US and was causing 60% of our client’s seed naming data issues. We helped by quickly building an application for correcting classified seed names, uploading and unifying seed data.

Team

QA
UI/UX
Project management
Product management
Front-end development
Back-end development

Build type

Hybrid

Tech stack

React
Typescript
GraphQL

What we did

Time restrictions

We had under 3 months to improve our client’s solution before peak usage (the start of planting season). We quickly updated and rearchitected a user interface flow to replace manual entry across multiple fields at once.

Improved UI
Quick wins for immediate value

International teams

This was an international effort with team members across 6 hours of time zones. Though this was challenging, we efficiently handled tests, subsequent analytics, and CI/CD requirements.

Automating deployment

The application was in a manual deployment state, but we were able to convert both repositories to fully CI/CD applications, evaluate the testing pyramid, and add the necessary unit, UI, and automation tests to achieve full CI/CD to launch before the planting season peaked.