Beck's Hybrids / Undergraduate Research
Overview
Becks' Hybrids is a family-owned seed company based in the United States, known for developing and distributing hybrid seeds and agricultural solutions. Becks specializes in providing farmers with innovative seed technologies to optimize crop performance and yield.
Roles
As a co-leader on this team I was responsible for direct communication with our client, where I led meetings and represented Purdue. Additionally, I was responsible for picking and carrying out design activities to direct the success of our team.
Goal
Beck's has become interested in genomic prediction models to streamline the development of corn hybrids, offering benefits like larger ears, taller plants, etc. Our team was tasked with researching breeders' needs and creating a dashboard for analyzing data from corn hybrids using predictive modeling capabilities.
Final Deliverable Overview
Final Dashboard Screens
Our team was able to deliver a series of high-fidelity, workable prototypes via Figma. Below is an overview of those screens, and annotations covering their functionality.
Disciplinary Knowledge
One of the most important parts of understanding our users, problem space, and Beck's as a stakeholder, was cultivating a real understanding of plant hyrbids, genomics, and breeding. Before joining this project I had no background knowledge in the agricultural space, this was a huge learning curve for myself and the team.
Important Terminology & Knowledge Domains:
Hybrid Plant Breeding and Genomics
Predictive Modeling
Data Visualization
Design Requirements
Beck's was able to provide our team with a list of constraints and feature requests. This list helped our team greatly reduce the complexity of our broader scope. To make this list more actionable, we used a prioritization activity to narrow down what was feasible for our team, and most valuable for Beck's.
Breeder Flow
In conjunction with our prioritization activity, our team was able to create a flowchart detailing what it looks like for our users to log into this software, take action, and make decisions.
Design Progression and Testing
Early Sketches and Low-Fidelity Mockups
As Beck's provided much of the validity and need for this specific solution and software, our team did not need to conduct secondary research or interviews into the efficacy of a breeding and predictions software. What our team did need to do was spend time ideating and generating UI solutions.
Constant Feedback & Testing
Throughout our research experience with Beck's we were in constant communication with our contacts; Our team received feedback from prospective users, subject matter experts, and we even had the opportunity to test our mid-fidelity designs with a Breeder.
From our testing we learned that having the ability to manipulate the generated visualizations would be a huge add for breeders. He suggested focusing more on projects and predictions rather than specific graphs. Retraining models is outside his expertise, so he suggested a notification system for tracking whether models are being expanded or newly created. For predictions and visualizations, he mentioned that he values flexibility in adjusting x and y axes, filtering data, and exporting both filtered and full datasets.
Our team took this data, refined our designs and delivered the following designs.
