MMehak Maini Mynzo Climate Dashboard

Case studyClimate data platform

From scattered data
to clear decisions

I designed the Mynzo Climate Dashboard end to end, from research to the front end: the live SaaS platform through which Mynzo delivers its services to clients, turning scattered field and satellite data into clear decisions.

My role
Product designerUX research to launch · front end
Team
Founder, product lead, tech team
01

Stage 1 of 5

The problem

Where we started, and what went wrong for clients.

Before the problem

Mynzo at a glance

01

What Mynzo does

  1. CO₂
    Companies plant forests

    to take carbon out of the air

  2. Mynzo looks after them

    We monitor, measure and report what they achieve

02

How Mynzo works

  1. 1 · Collect

    Our mobile app collects data on how each tree is growing.

  2. 2 · Analyse

    We run satellite analysis with our in-house AI models.

  3. 3 · Report

    We generate reports that show the results.

03

How a project is organised

Project Forest Boundary

Think of a school

  • Projectthe whole school
  • Forestone classroom
  • Boundaryone desk, where the work actually happens
How it unfolded

The story, from first version to today

  1. 1Built an MVP

    We had no users yet, so we shipped a first version early.

  2. 2Onboarded clients

    Real clients started using it, and we began watching how.

  3. 3Understood them

    Talked to users about what they needed and how they work.

  4. 4Found the gaps

    Four problems showed up in the MVP.

  5. 5Decided the platform

    Set what it needs and how it should work.

  6. 6Designed it

    New screens, built and released to clients.

What we noticed

The MVP, in four screens

We shipped an MVP early so clients could see their forests quickly. Once real clients were using it, I studied how they worked with it.

The four problems

What went wrong for clients, and for our team

  1. 01

    Built for everyone, so it worked for no one usability

    Too many clicks and too much to take in before anyone reached what they came for.

    In the MVPNew users needed a walkthrough from our team to find their own forest.

    How I solved this
    ???
  2. 02

    A display dashboard, not a decision dashboard decision

    The dashboard only showed numbers, so the decision was left to clients.

    In the MVPSurvivability showed 87.2%, but nothing said whether that was good or bad.

    How I solved this
    87.2% ?
  3. 03

    It couldn't keep up as data grew data

    Every new client added projects, forests and boundaries to one crowded screen.

    In the MVPLong lists and tables, with nothing to say where to look first.

    How I solved this
  4. 04

    Growth meant more manual work business

    Our team made reports by hand, trained every new user and answered questions about the numbers.

    In the MVPReports were files our team made and sent over, one by one.

    How I solved this
The challenge

How might we turn complex climate data into clear answers users can act on02, showing each person what matters to them03, on one platform that clients can use on their own01 and that scales as new clients join04?

My role

Product designer, end to end

I owned the experience from the first research session to launch, and built the entire front end myself.

I interviewed real users from client teams about the platform, to understand their role in the company, the tasks they do, the decisions they take, their goals and their pain points.

What I owned

  • User research
  • Information priority
  • Access model
  • Interaction and visual design
  • Front-end development

Who I worked with

  • FFounderProduct direction
  • PLProduct leadRequirements
  • TTTech teamBackend, AI and satellite
? USER RESEARCH INFORMATION PRIORITY ACCESS MODEL </> INTERFACE AND FRONT END
My approach

Four steps

  1. Step 1

    Understand the users

    Who uses the platform, and what do they need?

    Interviews · Information priority · Findings · User journey · Access control

    Jump to section →
  2. Step 2

    Decide how the platform works

    How do people move through it?

    Journey mapping · Findings · Decisions

    Jump to section →
  3. Step 3

    Design the screens

    What do users see and do?

    One template · Monitoring · Live demo

    Jump to section →
  4. Step 4

    Build and ship

    How did it go live?

    Front end · Hand-off · Release

    Jump to section →
02

Stage 2 of 5

Understand the users

Talk to the people who use it, and learn what each role needs.

Talked to users

Four types of users

I interviewed the users myself, one on one, to understand their role, their goals, their needs and the tasks they come to the dashboard to do.

Sustainability head

ESG head
Where do we stand, and what can I report?
Role
Reports the company's climate progress to leadership and the board
Goal
Know where every project stands, and share it with confidence
Tasks
  • Check project status
  • download the latest report
  • share it
  • flag problems to the team
Needs
Status, progress, alerts and ready reports. Rarely raw data or photos

What they assumeLeadership asks "where do we stand?", not "how many trees did we plant?"

Super admin

Platform owner
Where is each project, and who is responsible?
Role
Runs the platform for the company: projects, people and permissions
Goal
Every project has an owner and keeps moving
Tasks
  • Set up projects, forests and boundaries
  • assign managers
  • check activity
  • answer the auditor
Needs
Who owns what, what stage each project is in, latest activity, audit trail

What they assume"I don't need the data. I need to know someone is doing their job."

Field operations manager

EHSO
What data is still pending?
Role
Makes sure fieldwork happens on time, on the ground
Goal
Get every forest's data collected, on schedule
Tasks
  • Give the mobile app to field teams
  • track collection
  • follow up with whoever is behind
Needs
Stage of each forest, pending data, who is responsible

What they assume"If it's not done, whom do I call?"

Auditor

Independent verifier
Can I trust this data?
Role
Checks that every reported number is real
Goal
Verify the results before they are reported
Tasks
  • Check the project stage
  • review evidence
  • trace the audit trail
  • export data
Needs
Evidence, audit trail and completeness, not marketing language

What they assume"My job is to verify the outputs, not to consume them."

Information priority

What each role needs to see, and when

After studying each persona, their tasks and the actions they take, I listed the information each of them needs, and marked how important it is to each person: primary, secondary or tertiary.

Sust. headSuper adminField opsAuditorPrimarySecondaryTertiary

1Where are we?

Sust. headSuper adminField opsAuditor
Current project stage
Ground data collection status
Reports available
Latest project updates

2Who owns this?

Sust. headSuper adminField opsAuditor
Total projects
Total forests
Total boundaries
Project assignment
Assigned field ops manager

3Can this be verified?

Sust. headSuper adminField opsAuditor
Tree-level ground data
Satellite images

4What did we achieve?

Sust. headSuper adminField opsAuditor
Survivability
Sequestration
Trees planted
Saplings assessed
InformationSustainability headSuper adminField ops managerAuditor
Where are we?
Current project stage
Ground data collection status
Reports available
Latest project updates
Who owns this?
Total projects
Total forests
Total boundaries
Project assignment
Assigned field ops manager
Can this be verified?
Tree-level ground data
Satellite images
What did we achieve?
Survivability
Sequestration
Trees planted
Saplings assessed

PrimarySecondaryTertiary
User journeys

How each role uses the dashboard today

1 of 4

Super admin

Platform owner Goal: Every project has an owner and keeps moving

  1. Step 1Set upCreate projects, forests and boundaries
  2. Step 2AssignGive each project its managers
  3. Step 3CheckReview activity across projects
  4. Step 4AnswerRespond to the auditor

Needs to seeWho owns what, what stage each project is in, the latest activity, the audit trail

Field operations manager

EHSO Goal: Get every forest's data collected, on schedule

  1. Step 1Hand outGive the mobile app to field teams
  2. Step 2TrackFollow collection, forest by forest
  3. Step 3Follow upCall whoever is behind

Needs to seeThe stage of each forest, what data is pending, and who is responsible

Sustainability head

ESG head Goal: Know where every project stands, and share it with confidence

  1. Step 1CheckSee the status of each project
  2. Step 2DownloadGet the latest report
  3. Step 3ShareSend it to leadership
  4. Step 4FlagRaise problems with the team

Needs to seeStatus, progress, alerts and ready reports

Auditor

Independent verifier Goal: Verify the results before they are reported

  1. Step 1CheckConfirm the project stage
  2. Step 2ReviewGo through the evidence
  3. Step 3TraceFollow the audit trail
  4. Step 4ExportTake the data out

Needs to seeEvidence, the audit trail and completeness

Findings

What I found

Status didn't exist

Every role comes to the dashboard to know how each project is doing and which stage it has reached. The dashboard didn't show this anywhere.

So: Status became the first thing on every screen. Users first ask "is everything fine?", and only then "if not, why?", so the reasons sit one click away.

Each role has a different job

Each type of user needs to see and do different things on the platform.

So: I designed the whole access system around roles. A new user is simply given a role, with no setup for each person.

Task flow

Who does what, from setup to results

One flow runs across all the roles. Each person picks up where the last one left off.

Who
1 · Set up
2 · Give access
3 · Collect
4 · Process
5 · Monitor and act
Super admindashboard
SetupCreate the project, forests and boundaries. Upload existing data.
AccessGive dashboard access and assign a field operations manager
OverseeCheck every project is owned and moving
Field operations managerdashboard
App accessGive the mobile app to the field team
Track the teamSee who's done, who's behind, and follow up
Field teammobile app
FieldGo to the site and collect tree data
Mynzosystem
ProcessAI models and satellite imagery analyse the data
ESG head and auditordashboard
MonitoringSee status, results and reports; auditors verify the evidence
Access control

Who can do what

A person gets a role, and the role decides what they can see and do. I listed every function on the platform, grouped them into seven areas, and gave each one an owner and a reason.

No user is ever deleted. A user is only deactivated, so the work they contributed stays on record for audits.

FunctionOwnerReason
Create a dashboard user and give them a roleSuper adminOnly the super admin gives dashboard access
Create a partner app userField operations managerHe gives the mobile app to his field team. The app has a single role, so there is no role to choose
Change a dashboard user's roleSuper adminRoles exist only on the dashboard
Assign a user to a projectSuper adminAll assignments sit with the super admin
Deactivate a userSuper adminThis also covers removing someone from a project or resetting their access
Deactivate the super adminMynzoThe client emails Mynzo, and we deactivate the account
03

Stage 3 of 5

Decide how the platform works

Turn what we learned into the rules the platform runs on.

Scenarios

How I mapped every screen

I followed each user's journey: the steps they take on the dashboard to reach an answer and make a decision. Then I drew it out. The overview is the first thing everyone sees; each project opens from it.

Overview · All projectsFrom my FigJam board
Overview
(All projects)
1
Why these first?
These are the stages every project moves through, and stage status is what every persona cares about most. It is their top priority, and the first thing they look for.
Projects created · 05
Projects on track · 02
Projects with plantation data · 02
Projects with ground truth · 01
Projects under satellite processing · 00
Final reports · 00
DEFINE EACH STAGE: HOW?
Lowest / highest project stage?
2

Project status related

Projects created · 05
Projects ongoing · 02 (plantation data, ground truth and satellite processing)
Projects with no data · 03
Projects completed · 00 (all forests)
Reports available · 00
Total trees planted · 100
Records collected / planted · 10/100
Satellite-detected trees · 100
Total seq.
Total surv.

Overall

Action buttons
Timeline and duration? Will this happen every year? If yes, a year filter to be included.
First-level decisions are taken here, based on this data, which is to know if everything is going fine.
If things are not going well, then it is second-level info to check why.
  • Total trees planted · 100
  • Records collected / planted · 10/100
  • Satellite-detected trees · 100
  • Total seq.
How relevant is multi-selection for projects or forests? Why would a user want to view cumulative data across two projects / forests / boundaries?
Mental model: users are familiar with the Google Maps pattern of progressively revealing information as they zoom in. At higher levels they see broader regions, and as they zoom in, more detailed locations become visible.

A version I tried first

Project A2100 forests, 1100 in progress
Plantation data
500 forests
Ground data
400 forests
Satellite analysis
200 forests
Outcome generated
0 forests
Project B100 forests, 10 in progress
Plantation data
10 forests
Ground data
0 forests
Satellite analysis
0 forests
Outcome generated
0 forests
More cognitive load. Also, how would he know if these are going as expected?

What I chose instead

Project A2100 forests
On track2000
Project B500 forests
On track · 400Not on track · 10
One bar per project. He can tell at a glance whether it is going as expected.

Zoom in

Individual projectZoomed in from the overview
Individual
projects
★★

★ What we settled on

Project A

Overall

Forests on track · 10
Forests not on track · 10
Reports available · 02
Total trees planted · 100
Records collected / planted · 10/100
Satellite-detected trees · 100
Total seq.
Total surv.
Assigned manager's name
'On track' timeline to be decided forest-wise or boundary-wise? If forest-wise, it would be visible here. If only boundary-wise, then it will be visible at boundary level.
Recent updated / Recent milestonesmehak.maini
Time? Weekly? Monthly? Last opened (if possible). Depends on what frequency he logs in.
Updates are hierarchical, like:
  • 250 ground truth records collected for 17 forests. (Should these 17 be actionable?)
  • 2000 trees planted in 12 forests
Milestones achieved:
  • Ground data collection completed for 18 forests
  • Satellite outputs available for 12 forests

Project A status

Plantation data500/500 forests
Ground data400/500 forests
Satellite analysis200/500 forests
Final reports0 forests
  • Satellite-detected trees? Trees identified?
  • Canopy cover %
  • Seq.
  • Survivability
Trees assessed: trees that were selected / evaluated as part of an assessment process (growth records), not trees detected after satellite analysis.
Mental model: users are familiar with the Google Maps pattern of progressively revealing information as they zoom in. At higher levels they see broader regions, and as they zoom in, more detailed locations become visible.
On Track ✅

Not On Track ❌

Details (date-wise, display only boundary (forest-wise))

Project A status

Plantation data500/500 forests
Ground data400/500 forests
Satellite analysis200/500 forests
Outcome generated0 forests
  • No. of records collected / planted
  • Canopy cover %
  • Seq.
  • Survivability

🌳 🌱 🛰️ 🥜

Zoom in

Zoom in

Tap

Zoom in

Scroll sideways on small screens to see the whole board. Project names are placeholders.

Findings and decisions

What the boards showed me, and what I decided

01

On track or off track

What I found

Users asked for status. What they didn't know was that they needed on track and off track.

What I decided

I introduced on track and off track. If even one boundary is off track, the whole forest is off track, so a delay is never averaged away.

See it in the product
02

Patterns users already know

What I found

The data is complex, and this was new software for our users.

What I decided

The map works like Google Maps and reveals more as you zoom in. Detailed data works like an Excel sheet.

See it in the product
03

A lightweight monitoring screen

What I found

Monitoring is where users decide. The detail they need to act lives on other screens.

What I decided

I kept monitoring light and linked it to the rest of the dashboard. The super admin sees a project is off track, taps its manager, and opens the access page. There she finds why: 500 of the expected 1,200 tree samples are collected.

04

No random grouping

What I found

The old dashboard let users select any mix of projects, forests or boundaries. No client needs to see two unrelated projects together.

What I decided

I removed multi-selection. Every view follows the grouping that already exists: project, forest, boundary.

05

Reni, a helping AI

What I found

Users need help in different ways: reports, quick answers about their data, or getting in touch with our team.

What I decided

Reni, our AI assistant, helps with all of it: it generates reports, answers anything about the data, and connects clients with the Mynzo team. If a client ever needs a metric across projects, Reni handles it.

See Reni in the product
06

When there is no forest level

What I found

Some clients go straight from project to boundary, with no forests in between.

What I decided

I proposed allowing two levels. The team kept three because of the backend, so I designed a zoom from project straight to boundary.

07

A year filter built from the data

What I found

Every client collects data on their own schedule: yearly, quarterly or once in two years.

What I decided

The filter shows only the years we have data for. If a client has data for 2023 and 2026, those are the two options.

See it in the product
08

What I left out: recent updates

What I found

My board had a recent updates panel. But users first check on track or off track, and only then look deeper.

What I decided

I kept it off the monitoring screen, so the first step stays clear.

04

Stage 4 of 5

Design the screens

Give every decision a screen that clients can read at a glance.

Screens

Why the map leads the screen

Monitoring is where clients spend most of their time. Every project has forests, and every forest is split into mapped plots called boundaries. Three reasons made the map the centre of it.

  1. 01

    It matches how users already think

    Everyone has used Google Maps. Zooming from a country to a street is a pattern they already know, so moving from a project to a boundary needs no training.

  2. 02

    Too much data for one glance

    Every new client adds more projects, forests and boundaries. A table of all of them is unreadable; a map holds everything and lets users look closer only where they need to.

  3. 03

    Every number belongs to a place

    Field data is collected inside a boundary. Showing it on the map tells users not just what is happening, but where.

Live demo

See the decisions in the real platform

Scroll on and the dashboard opens full screen. Every view carries the reason behind it.

Loading the live dashboard…

Walkthrough

Nine screens, from all projects to a single tree

Scroll through the live product. Each ?Why on a screen opens the reason behind that decision.

01 / 09 · All projects

Is everything fine?

WhyEvery user opens the dashboard with the same first question. So the overview answers it with status, before any number. Finding: status didn't exist ↑

Monitoring, all projects: six projects on a satellite map, status filters at the top and five headline metrics along the bottomStatus comes first. On track or off track is what every user looks for, so the counts sit at the top and filter the map in one tap.A problem shows without a click. Every project carries its status on the map, so nobody has to open a project to find out.For convenience. Users can search any boundary directly, without first choosing a project and a forest.Clients collect data on their own schedule. The date filter offers only the years that actually have data.Results follow status. Area, carbon and survivability are the outcome of all the work, so they sit below the map, not above it.

02 / 09 · Project

One project, one level down

WhyGrouping follows the structure clients already work in: project, forest, boundary. No random mix of places, so every number on screen belongs together. Decision: no random grouping ↑

Monitoring, one project: its forests on the map, forests on track and not on track at the top, and the project's metrics belowThe same question, one level down. Forests on track and not on track let users follow a problem to where it is.Only this project's forests. The map shows nothing from other projects, so there is nothing to mix up.Numbers follow the map. The totals change to this project, so what users read always matches what they see.

03 / 09 · Forest

A delay is never averaged away

WhyIf even one boundary is off track, the whole forest is off track. A problem in one plot can't hide inside a good average. Decision: on track or off track ↑

Monitoring, one forest: five boundaries on the map, two off track in red and three on track in greenOff track stands out. Off-track boundaries are red and flagged, so the eye goes to them first.On track is marked too, with a tick, so status never depends on colour alone.Counts before detail. Users see how many boundaries need attention before they look at any one of them.

04 / 09 · Boundary

Where the data actually lives

WhyEvery number is collected inside a boundary. So this is where all the data comes together, as layers on one map. Research: information priority ↑

Monitoring, one boundary: Sanjay Van outlined on satellite imagery, with the Layer Panel open on the left and the boundary's metrics belowGrouped the way clients track their work: metrics, plantation, flora and fauna. I took the groups from clients' own data sheets.Nothing looks broken. Layers without data yet stay visible but disabled, so clients know what's coming.The real plot. Data is collected inside this boundary, so the map shows exactly where it comes from.

05 / 09 · Land cover

Land cover to compare against

WhyLand cover is the layer clients compare everything else against: how much planting happened in each type of land, and how many species sightings or disturbances fall in it.

The Land Cover layer on: the boundary coloured by forest, water, grass, shrub and bare land, with land-type metrics belowThe base for comparison. Clients turn land cover on to read every other layer against it.Every colour is named in the legend, so no one has to guess what the map shows.The numbers switch with the layer: hectares of each land type, for this boundary.

06 / 09 · Tree markers

See where the work was done

WhyThe super admin and the EHSO need to know the field team really went to the site and did the work as expected. Research: information priority ↑

Tree Sampling layer on: a marker for each sampled tree inside the boundary, with sampling metrics belowProof of presence. Each marker sits at the exact latitude and longitude where its photos were taken, so coverage and gaps show on the map.How far collection has got: trees sampled, sampling percentage, species and geo-tagged images, at a glance.

07 / 09 · Tree record

The evidence behind every number

WhyEvery role needs proof, not only the auditor. The super admin and the EHSO check that teams measured each tree as expected; the auditor verifies the numbers. Research: information priority ↑

A tree sampling record: species, date, girth and height, GPS coordinates, and photos of the tree's circumference, height and leafA number you can check. The photo shows the measurement itself, so no one has to take it on trust.Tied to one place. The coordinates link the record back to its marker on the map.

08 / 09 · Species

Biodiversity, with honest data

WhyBiodiversity goes beyond trees: plants, animals, photos, videos and voice notes. Saying how certain each record is keeps the data trustworthy enough to audit.

A species sighting of an Indian Peafowl with its confidence level, scientific name, approximate location and photoCertainty is stated, not assumed. Each sighting shows its confidence level.No false precision. When the survey has no GPS per sighting, the location is labelled approximate.Counted by group: birds, mammals, reptiles and amphibians, so a gap in the survey is easy to spot.

09 / 09 · Reni

A helping AI, on every screen

WhyReni is there to help, with anything. It answers questions about the data, generates and exports reports, and helps clients get in touch with the Mynzo team. Decision: Reni, a helping AI ↑

Reni, the AI assistant, open beside the map, answering with a year-by-year carbon table and offering to download itHelp without leaving the screen. Reni opens beside the map and knows which site the user is on.Answers in the user's own data, not generic text.A question can become a report. Reni offers to download the answer as a file.
Problems solved

The four problems, and how I solved them

01

The problem · usability problem

Built for everyone, so it worked for no one

How I solved it

One template, shaped by each role. The information priority from my research decides what each person sees first, access is set by role, and the map works like Google Maps, so nobody needs a walkthrough to find their own forest.

02

The problem · decision problem

A display dashboard, not a decision dashboard

How I solved it

Status comes first. Every project, forest and boundary says whether it is on track or off track, and if one boundary is off track, the forest is too. The question “is everything fine?” is answered before any number, and the reason is one click away.

03

The problem · data problem

It couldn't keep up as data grew

How I solved it

The map holds everything and reveals more as users zoom in: project, then forest, then boundary. Every view follows that grouping, the numbers change with the level, and search finds any boundary directly.

04

The problem · business problem

Growth meant more manual work

How I solved it

Clients now serve themselves. Every report lives in one library and clients can build their own, field data comes in through a template the system reads automatically, and access is managed in the platform instead of by our team.

Screen 02 · Reports page

The decisions behind Reports

In the MVP, reports were files to download, and our team made each one by hand. On the new platform, every report lives in one library, and clients can generate their own.

Walkthrough

Six screens, from the library to a custom report

Scroll through the live product. Each ?Why on a screen opens the reason behind that decision.

01 / 06 · Library

Every report in one place

WhyIn the MVP, reports were files our team made by hand and sent over. Reports were a primary need for the sustainability head and the auditor, so they get one library they can use on their own. Problem 04: growth meant more manual work ↑

The Reports page: tabs for plantation and biodiversity reports, a table of reports with project, forest, boundary, date, status and download, and a Generate Custom Report buttonTwo kinds of report, two tabs: plantation and biodiversity.Every report is tagged with its project, forest and boundary, the same spine as the rest of the dashboard.Status and download sit side by side. “Ready” means one click to download.Generate Custom Report lets clients build their own, without asking our team.

02 / 06 · Filters

Filter where you're already looking

WhyWith many projects, forests and boundaries, a long list of reports is only useful if you can narrow it fast. Filtering in the column keeps the table and the filter in one place.

The project filter open in the Project column header, with a search box and a checkbox list of projectsEach column filters itself: project, forest, boundary, date and status, with search inside the list.Counts link through: “2 Forests” or “3 Boundaries” show what a report covers.

03 / 06 · Steps

One step at a time

WhyA report builder can easily become a wall of options. Numbered steps that unlock in order make it hard to build a wrong report.

The Generate Custom Report dialog: step 1 select project, step 2 select forests, step 3 select fields, locked until a project is chosenStep 1: choose a project. Everything else depends on it.Step 2: choose forests. It defaults to all forests, so the common case needs no extra clicks.Later steps stay locked with a clear note until the step before is done.

04 / 06 · Fields

Fields in plain language

WhyThe people building reports are not data specialists. Plain descriptions mean they pick the right columns the first time.

Step 3 of the custom report: a list of data fields such as plantation count, plantation date, species, tree count and sequestration, each with a one-line descriptionEvery field explains itself in one line, like “Number of plants or trees planted per forest or project”.Units are stated, like sequestration in tCO₂e, so the numbers can't be misread.

05 / 06 · Status

Never wonder if it worked

WhyGenerating a report takes time. A visible status shows the request was received, so nobody clicks twice or emails our team to ask.

The report list with the new custom report at the top, marked In Progress, while the other reports are marked ReadyThe new report appears at the top of the list straight away.“In progress” turns to “Ready” when the file is done, and the download button switches on.

06 / 06 · Biodiversity

Reports in the frameworks clients use

WhyClients report biodiversity against these frameworks. Labelling each file by framework means they can find the one they need to submit.

The Biodiversity Reports tab: framework documents such as GRI 101, TNFD and IUCN/BIRS for each boundary, with project, forest, boundary, framework, date and downloadGenerated from a boundary's biodiversity analysis and named for the framework: GRI, IUCN/BIRS, No Net Loss, IFC PS6, TNFD.A framework column shows at a glance which standard each file follows.
05

Stage 5 of 5

Build and ship

Build the front end and release it to clients.

Front end

I built what I designed

I built the entire front end myself, in HTML, CSS and JavaScript, so what shipped matched the design. Because I handed over working code, not mock-ups, our developer could ship it almost straight away. The tech team connected it to the backend, our AI models and satellite processing.

  1. DesignEvery screen and its states
  2. BuildThe full front end, by me
  3. ConnectBackend, AI models and satellite data, by the tech team
  4. ReleaseReleased to clients
Client response

What clients said in the demo

Said during the demo by clients who went on to sign. Four more clients came on board in October, after seeing the updated dashboard.

It looks very easy to use.
Client, during the demo
It caters to all our needs.
Client, during the demo

They also singled out the roles, and the access given to each role, as exactly what they needed.

Who uses it now

  • DDADelhi Development Authority
  • IOCLIndian Oil Corporation
  • A leading steel manufacturerIndia
  • A leading construction companyIndia
  • A petroleum companyMalaysia and Singapore
  • A school groupIndonesia, with plantations across its branches
Reflection

What I learned

  • Start with people, not data

    The answer each person needs matters more than how much data we can show.

  • Take rough work to review early

    Changes on paper cost far less than changes in code.

  • Write down what you leave out

    The reason something isn't on a screen matters as much as what is.