Operational Analytics

Operational Analytics

Operational Analytics

Transforming Raw Data into Actionable Insights

Transforming Raw Data into Actionable Insights

Role

Lead Designer
Product Execution Lead

Platform

WingspanAI Web

Teams

Design
Product
Engr/QA

Design, Product, Engr/QA

Goodbye, Manual Math

Goodbye, Manual Math

Data Should Work For You

Data Should Work For You

Data Should Work For You

WingspanAI promises to be a remote control center for farm operations, offering advanced data solutions to increase efficiency and profitability. However, its usage was at an all-time low, revealing a large disconnect in how we highlighted data. Farm managers needed specific metrics and big-picture insights across large time-scales, not granular drive sessions with excessive manual consolidation.

After many iterations and technical challenges, we aggregated operational data into an interactive, comprehensive view with built-in progressive disclosure, making it easy for users to evaluate their overall farm's performance before drilling in for greater details.

Say hello 👋🏼 to the new and improved Operational Analytics.

We delivered a completely redesigned Operational Analytics at the end of Q3, doubling WingspanAI Web usage and dominated as the most used tool in the entire app. We were able to shift the purpose of WingspanAI Web from live monitoring to post-operational review.

Follow along to see how we got here.

The Problem

How Many Clicks Does It Take To Get To The Center?

How Many Clicks Does It Take To Get To The Center?

How Many Clicks Does It Take To Get To The Center?

To deliver real value with a smart, data-driven tractor, we had to make insights accessible. Unfortunately, simple questions like how much a user saved by choosing an electric tractor over diesel became a scavenger hunt through every drive session with manual calculations at the end. Let's be honest, no one wants to crunch numbers after a long day in the field.

USER PROBLEM

User feedback revealed that users relied on multiple sources to gather information and manually consolidated the data themselves.

HYPOTHESIS

We believe that users want a single, easily accessible source that delivers valuable, actionable data immediately for more effective reporting and future operations planning.

1
BUSINESS PROBLEM

Low engagement and user complaints around the promise of a "data-driven" product.

HYPOTHESIS

We believe that designing an interactive dashboard that delivers better analytics will drive engagement and showcase fleet performance effectively.

2

Priorities, Pains, and Research

Priorities, Pains, and Research

Priorities, Pains, and Research

Different Fields, Different Needs

Different Fields, Different Needs

Different Fields, Different Needs

There were distinct differences between our top user segments, dairies and vineyards, like their operating hours, key data points, and size. We prioritized the larger vineyard user base, while designing a scalable solution that could easily adapt to dairy users in the future.

To uncover the specific challenges our users faced with their tractor data today, I directly spoke with crew supervisors and farm managers.

Fragmented Data

Reconstructing a complete day's work from individual drive sessions was time-consuming and frustrating, especially for fleet-wide weekly reviews.

No Separation by Operation Type

Mowing will be evaluated differently than spraying, and managers couldn't differentiate between the drive sessions.

Missing Key Metrics

Essential Metrics like Acres Covered and Efficiency were missing, leaving users blind to the performance of their fleet.

Lack of Operator Tracking

Without tools to track operator performance, farm managers couldn't assess readiness for assist or automate modes.

Hard to Review Incidents

Incidents were reported independently of this feature, causing users to navigate back and forth.

No Savings Trends

The inability to track and compare savings prevented users from assessing the value of their electric fleet.

To deliver an intuitive and impactful user experience, I evaluated other analytics powerhouse tools and defined core design principles. This research guided my design process while maintaining ease of access and usability at the center of all the complexity and chaos.

Google Analytics

Google Analytics

Google Analytics

Tableau

Tableau

Tableau

PowerBI

PowerBI

PowerBI

Create custom dashoards

Free-Select Dates or Custom Date Range

Variety of Data Visualizations

Easy-to-use Search Tool

Tooltips for Statistics

Featured Insights

Exportable PDFs

Manipulate & Customize Databases

1

Minimize Information Overload

Present data in digestible, prioritized layers to help users focus on the most important information

2

Clear Data Visualizations

Use simple, effective visual elements to make complex metrics easy to understand at a glance

3

Logical Information Funnel

Organize content to guide users from high-level summaries to more granular details

4

Map-Centric Design

Highlight interactive maps that provide a holistic view of farm operations and context for incident reviews

Defining The Solution

Ask, Then Answer

Ask, Then Answer

Ask, Then Answer

Our goal was to support better review and future operation planning for our users by giving meaningful, actionable data. We started by defining user stories in the form of real questions farm managers needed answers to, which shaped the core functionalities and new metrics to highlight.

What work was done on my farm and where?

How was the quality of the work performed on my farm (coverage, efficiency)?

Were there any issues during my operations?

How well did my operators perform (idle time, efficiency)?

How much have I saved using MK-Vs over diesel tractors?

What is the expected run time of my fleet for different operation types?

Scalable Data Up To a Month

Aggregated Stats by Op Type

New Metrics like Acre Coverage

Savings Trends by Month

Incident Reviews with Context

Filter and Search Operations

Track Operator Performance

Video Playback of Operations

Battery Runtime By Op Type

To provide users with these answers, we needed to rethink how we bundled and displayed their data.

Technical Challenge

Transforming Fragmented Drive Sessions Into One View

Transforming Fragmented Drive Sessions Into One View

Transforming Fragmented Drive Sessions Into One View

We noticed that different operations had different performance expectations. For example, with spraying, speed played a large role in determining the efficiency of the operations. If the operator drove too fast, all the input would be lost in the air and not enough would cover the vines, reducing the efficacy of the product.

The challenge was that our tractor only reported data in individual drive sessions, which were often too short to give users a clear picture of the day's work. How could we provide a comprehensive view of the farm across all operations without forcing users to sift through fragmented sessions?

Each implement has a unique type, like mower, which correlates to an operation type. We grouped all drive sessions with the same tractor-implement pair into a daily summary, accounting for breaks and handoffs. These daily summaries were then organized by operation type, enabling users to analyze trends over weeks or months.

In addition to the technical challenge of transforming data, we also developed new methodologies to capture and calculate other missing metrics. This approach brought us much closer to providing our users with actionable, accurate insights that answered their open questions.

The Solution

The Solution

The Solution

Piecing It All Together

Piecing It All Together

Piecing It All Together

This project was a journey of continuous iteration, with many feedback loops and adjustments to ensure the final product significantly improved our users' review and planning processes. From gathering the right data and aggregating it effectively, to presenting it in an intuitive, non-overwhelming way, every piece had to come together. Through progressive disclosure and methods like A/B testing, I created a design so that users could easily analyze their farm's performance, both from a high-level monthly view or down to a specific moment in time.

Dashboard

Dashboard

To effectively communicate the scale of this redesign, users received an introductory guide highlighting the key metrics they can now access from this new analytics tool. The landing page consolidates all work performed by the MK-V fleet into a single, interactive dashboard so users can quickly assess the performance and efficiency of their farm.

To effectively communicate the scale of this redesign, users received an introductory guide highlighting the key metrics they can now access from this new analytics tool. The landing page consolidates all work performed by the MK-V fleet into a single, interactive dashboard so users can quickly assess the performance and efficiency of their farm.

Operational Review

Operational Review

By selecting an operation, users will be brought to a focused view of that type of work across their MK-V fleet. Here they can understand how power draw impacts performance, plan charging schedules, view coverage, and take a look at all the tractor-implement daily summaries to further analyze.

By selecting an operation, users will be brought to a focused view of that type of work across their MK-V fleet. Here they can understand how power draw impacts performance, plan charging schedules, view coverage, and take a look at all the tractor-implement daily summaries to further analyze.

Daily Summaries

Daily Summaries

We aggregated all drive sessions by tractor and implement pair across a single day. The joint video player-map feature allows users to go back in time and replay the work. They can hone in on specific camera angles and assess any hazards or safety incidents that may have occurred.

We aggregated all drive sessions by tractor and implement pair across a single day. The joint video player-map feature allows users to go back in time and replay the work. They can hone in on specific camera angles and assess any hazards or safety incidents that may have occurred.

Editing In Post

Editing In Post

In case operators forgot to select an implement at the start of their drive session, or selected the wrong one, we enabled users to be able to edit the implement for any drive session after it was uploaded to the analytics tool. There, we would dynamically reconfigure aggregations so the analytics tool stayed as up-to-date as possible.

In case operators forgot to select an implement at the start of their drive session, or selected the wrong one, we enabled users to be able to edit the implement for any drive session after it was uploaded to the analytics tool. There, we would dynamically reconfigure aggregations so the analytics tool stayed as up-to-date as possible.

Operator Reviews

Operator Reviews

We included a brief look into operators based on their unique login codes ont he tractors. This was especially beneficial to determine star performers, those who required more training, and those who were ready for assist and automate features.

We included a brief look into operators based on their unique login codes ont he tractors. This was especially beneficial to determine star performers, those who required more training, and those who were ready for assist and automate features.

Search

Search

To enable a free path secondary to the main purpose of historical operational reviews, we developed a search feature that would highlight specific daily summaries. We also included filters so users could quickly answer questions like "which operations had over 10 safety incidents" and deduce patterns from their findings.

To enable a free path secondary to the main purpose of historical operational reviews, we developed a search feature that would highlight specific daily summaries. We also included filters so users could quickly answer questions like "which operations had over 10 safety incidents" and deduce patterns from their findings.

Retrospective

Retrospective

Retrospective

Rome Wasn't Built In A Day

Rome Wasn't Built In A Day

Rome Wasn't Built In A Day

We completely transformed how users engaged with WingspanAI, from real-time monitoring to post-op insights. With easily accessible metrics, expandable incident reviews, and aggregate statistics, our users felt like they could truly wield the power of a dynamic, data-driven fleet at their fingertips.

However, this project took many months of iteration, feedback, R&D, and testing before we delivered it to our users. We faced many technical constraints along the way, but our collaborative teamwork helped us overcome all of the hurdles along the way. I grew a lot as a designer, learning to adjust on-the-fly to constraints and advocate for our work alongside Product and Engineering.

C

O

M

E

O

S

A

Y

O

O

H

E

L

L

0

This feels like a green square type of moment. Let's Chat!

© 2025 Manjari Maheshwari

C

O

M

E

O

S

A

Y

O

O

H

E

L

L

0

This feels like a green square type of moment. Let's Chat!

© 2025 Manjari Maheshwari

C

O

M

E

O

S

A

Y

O

O

H

E

L

L

0

This feels like a green square type of moment. Let's Chat!

© 2025 Manjari Maheshwari