IBM — Watson Analytics, Carbon System, and SPSS Modeler redesign
Product Designer/Front‑end across Watson Analytics and Data Platform; SPSS Modeler redesign, AP design system, early Watson Analytics.
Team make-up
Collaborative work with engineers, PMs, and data scientists at IBM. Main design lead for this area of work. Management/leadership work w/ newer designers and cross-org communication.
Overview
This set of work focused on the interaction and visual redesign of SPSS Modeler — data scientists can create flows to discover, prepare and blend data, develop and manage models, and visualize the results. No coding is required.
Approach
Initial discovery and research; connect with stakeholders and cross-collaborators as visuals were made; confirm throughout and at the end with research and feedback mechanisms.
Understanding the space
Determined the full corpus of charts to overhaul, cataloged usage, and interviewed users to uncover edge cases.
Competitive audit
Looked across the field for must‑haves and “nice to haves”; discoveries included mini‑maps, deeper drill‑ins, and motion helping comprehension.
Wireframes and prototypes
Shared wireframes with stakeholders and built coded prototypes to explore responsive behavior. Visual experimentation surfaced that users primarily sought the key number; later designs emphasized a clearer data‑input region.
Finalization
Iterations converged to match and connect with the Analytics Platform design language. Built coded prototypes of specific interactions; worked closely with dev on edge cases (too few/too many values), tooltips, help states, and clarity affordances.
Advanced work
Progressed to more advanced chart types (e.g., neural networks) and deeper interaction patterns. Predictor chart learnings informed brusher concepts to expose full value ranges adjacent to tables.
Outcome
Patterns from this redesign still appear in product checks years later.