Arjun TS

SotaOG

Months of oilfield data, readable at a glance.

SotaOG is an analytics platform for oil and gas operators. Their data lived in SCADA screens, paper tank tickets and separate finance systems. I designed the layer that brings it together into one view a team can act on.

Role
Lead Product Designer, founding designer
Team
Product, Strategy, Growth
When
February 2022 to April 2023
Status
Shipped
SotaOG truck ticket reconciliation with the history bar

−64%

Dashboard task time

8.3 to 3 minutes

+52%

Task completion efficiency

across core workflows

−37.2%

Reconciliation time

for tank tickets

  • 2 more industries, manufacturing and ESG, later adopted the platform.

The problem

Oil and gas teams run on real-time accuracy. Their data was static, scattered and partly on paper.

  • Decisions were delayed and anomalies were missed.
  • Reconciling tickets by hand leaked revenue.
  • It was hard to see how production had behaved over months.

SCADA is the supervisory software that monitors field equipment. Operators know it deeply, so the aim was to build on that familiarity, not replace it.

My role

  • End-to-end UX, information architecture and data visualisation.
  • The 3D digital twin, modelled in Blender and built with Three.js.
  • Workshops with the CEO, petroleum engineers and finance leads.
  • A small, enterprise-ready design system.
  • Mentoring junior designers, and keeping design aligned with engineering, data and the business.

Principles

  • Speed of understanding matters more than speed of clicking.
  • Summarise first, reveal detail on demand.
  • Keep what operators already know from SCADA.
  • Order every screen by the operator’s workflow.

Design logic

Design logic map for SotaOG
Mapping the operator’s path through the data

Four solutions

1 of 4

A dashboard that says what changed

BeforePeople couldn’t quickly read how production was behaving, or where attention was needed.

ChangeData grouped by relevance instead of the raw SCADA hierarchy. Every view answers the same three questions in order: what changed, why, and what to do next.

ResultFaster interpretation, and quicker team response.

Annotated optimisation dashboard: well health, insights in words, and ROI by block

2 of 4

The History Bar

BeforeThousands of sensor readings made mismatches and production drops almost impossible to spot by hand.

ChangeOne scrollable strip compresses months of data. Colour marks each day as matched, missing, in error or a sensor fault, and any day opens into its detail.

ResultMismatched days are spotted at a glance, not found by sorting.

Annotated truck tickets screen with the History Bar

3 of 4

Tickets that reconcile themselves

BeforeMatching tank tickets to run statements by hand was slow and financially risky, especially when tickets were lost or illegible.

ChangeTickets are read with OCR and matched against run statements and sensor readings. A drop in tank level now creates an auto ticket on its own, and bad readings are flagged instead of billed. Mismatches are fixed in the table, and everything stays searchable.

ResultReconciliation time down 37.2%, and payments for thousands of barrels that had been missed were reconciled once auto tickets went live.

Annotated reconciliation table comparing auto tickets, ticket images and run statements
Three sources, compared on one row.
Annotated tank management screen with corrected tank levels and auto tickets
Auto tickets, created from tank drops.
1 / 2

4 of 4

A digital twin of the facility

BeforeSCADA diagrams have no sense of space, which made troubleshooting slow.

ChangeA 3D model of the facility with live data on equipment and valves, colour-coded alerts for flow problems, faulty sensors and low tanks, and drill-down into any sensor.

ResultEngineers preferred it to static SCADA panels, and found faults faster.

Annotated real-time analytics screen with the 3D model and ranked alerts

One missing ticket, end to end

Every feature on its own looks like a dashboard. The value shows when you follow one problem through the team. This is a composite journey, built from our workshops with operators, petroleum engineers and finance leads: a single tank ticket that never arrived.

  1. 1

    Operations manager

    Is anything wrong today?

    Is anything wrong today?
    Before
    Log in to SCADA, export to a spreadsheet, scan the rows.
    With SotaOG
    Well health and the revenue at stake on one screen, with insights saying where to look.
  2. 2

    Revenue accountant

    Which days don’t match?

    Which days don’t match?
    Before
    Sort and filter thousands of ticket rows, hoping the gap shows.
    With SotaOG
    The History Bar marks every missing day across months. One switch puts the problems first.
  3. 3

    Production engineer

    Was that oil really sold?

    Was that oil really sold?
    Before
    Hold tank gauges against paper tickets by hand.
    With SotaOG
    The tank chart flags the bad reading and states that no ticket was billed for it.
  4. 4

    Field technician

    Where is the faulty sensor?

    Where is the faulty sensor?
    Before
    Trace a flat SCADA schematic, then walk the site to find it.
    With SotaOG
    The digital twin pins the tank with its live readings, and the alert is ranked by severity.
  5. 5

    Revenue accountant

    Can we close the month?

    Can we close the month?
    Before
    Match tickets to run statements one by one, often long after the fact.
    With SotaOG
    Auto ticket, ticket image and run statement sit on one row, and the gap is fixed in place.
  6. 6

    Finance lead

    What will we earn?

    What will we earn?
    Before
    Forecast in a separate finance system, blind to what happened in the field.
    With SotaOG
    Forecasts run on the same, now-clean data, with each model’s fit shown beside it.

Before SCADA, paper tickets, spreadsheets and a separate finance system, each holding part of the truth.With SotaOG One record the whole team works from, from the first alert to the forecast.

Getting the twin right

It took four versions. Each one fixed what the last one got wrong, and I modelled the final facility in Blender and built it in Three.js myself, because the team had no Three.js developer.

Four versions of the facility view, from a flat SCADA-style plan to the digital twin
From a flat plan to the real site.

Across the platform

The same rules ran through every view: the summary first, colour only for state, and detail on hover. These are the screens operators, engineers and finance leads used every day.

Annotated depletion strategy chart for every well
Which wells are running out, and how soon.
Annotated production forecasting screen with three decline-curve fits
Forecasting production, with the model in the engineer’s hands.
Annotated optimisation opportunities treemap
Where the money is, sized by value.
Annotated well status timeline with an alert
A well’s day, with the alert explained.
Annotated compressor downtime chart
Downtime by compressor, and why.
Annotated hourly efficiency chart for every well
Every well, hour by hour.
Annotated pipeline network map
The pipeline network, with its alerts.
1 / 7

Details that did the work

Operators, engineers and finance leads only act on data they trust. These details earn that trust.

Four details from SotaOG: confidence on every forecast fit, a bad reading that explains itself, thresholds written into the legend, and labels placed on the plot

Minimal isn’t always clear

The operational envelope shows the range in which a well runs most efficiently under many limits at once. My first plots either showed everything or stripped so much away that nothing could be acted on. Working through the data with the team, we found that only a few variables decide the answer.

In complex domains, clarity means the right depth, not the least.

Four earlier attempts at the operational envelope plot
Four attempts that didn’t work.

The final plot answers one question first: how much the well can produce, and what is holding it back. Everything else supports that answer.

Annotated final operational envelope plot
The plot that did.
The plot in use.
1 / 2

What people said

  • Technicians found issues faster.
  • Finance leads trusted their cash-flow predictions more.
  • Management saw shorter decision cycles and fewer surprises.

Reflection

In oil and gas, design isn’t decoration. It is clarity, reliability and precision under pressure.

This project shaped how I think about enterprise data. The platform later extended to manufacturing and ESG reporting.

Next: Concepts.schoolEngineering mathematics you can see and touch.