← All projects

02 / INTERACTIVE SOFTWARE · BUSINESS TOOLS

Indy.

From local data to a community intelligence report.

  • Python
  • React.js
  • PostgreSQL
  • API integration
  • Data integration
Explore the project ↓
Supporter records underpinning the platform
≈14,000
Requirements, deployment and staff training
End to end

Approximate project scale from my CV; the portfolio demo uses fictional data.

CONTEXT

Professional experience · Community intelligence

MY ROLE

Application development and reporting workflows

ON THIS PAGE

01 / THE QUESTION

Bring a place into focus.

Understanding the community around a proposed development means bringing together different kinds of information: potential supporters, local concerns, geography and political context. The challenge is to turn that information into a report that helps people plan engagement.

My contribution

My work on Indy involved application development, API integration and community intelligence reporting workflows. The report brings an executive summary, potential supporter profiles, a local support map, known objector groups, support themes, political context and resident quotations into one browser-based experience, followed by a summary of what those findings mean.

Inside the report

Start with the overall picture

An executive summary brings the support opportunity, likelihood of organised opposition and a key insight together so readers can orient themselves quickly.

Explore who, where and why

Potential supporter profiles and a local area map help explain where support may come from. Support themes describe the issues that could motivate engagement.

Understand the surrounding context

Known objector groups sit alongside council composition and MP information. Resident quotations add qualitative evidence to the broader picture.

Connect findings to a next step

The closing section draws the findings together and sets out a path from further research to supporter mobilisation and planning-ready evidence.

HOW IT FITS TOGETHER

From data to a useful workflow.

A simplified view of the original application’s data flow. Vercel hosts the frontend; Railway hosts the backend and PostgreSQL. The portfolio walkthrough runs independently with fictional data.

  1. 01

    Bring sources together

    Records · external APIs

    Combine supporter records and local context.

  2. 02

    Prepare & store

    Python · FastAPI · PostgreSQL

    Clean, organise and retrieve location-specific information.

  3. 03

    Request & explore

    React.js interface

    Choose a location and navigate the reporting experience.

  4. 04

    Deliver the report

    Backend reporting workflow

    Present summaries, mapping, themes and political context.

02 / GET A FEEL FOR IT

Every insight should have a trail.

The full report combines summaries, maps, profiles and local context. This smaller, independent demo focuses on one reporting principle: choose a fictional place and follow an insight back to its invented source and limitations.

INSIGHT → SOURCEFictional data
Select an insight to follow its source.
BEHIND THE INSIGHT

Fictional community questionnaire

64%
What it says
77 of 120 fictional respondents selected transport. The displayed percentage is rounded.
Scope
120 invented responses. Multiple-choice question about local priorities.
What it cannot tell us
This is an illustrative respondent sample, not a representative estimate of residents.
Illustrative source · Not real evidence
All places, responses and sources are invented for this walkthrough.

03 / HOW I APPROACHED IT

The decisions
behind the work.

01Layer the report for different levels of detail

The executive summary gives a quick overview, while profiles, mapping, themes and political context let readers explore the detail relevant to their question.

02Bring different sources into one workflow

API integration and data integration support a report that combines geographic, demographic and local context. Presenting these together helps readers connect information that would otherwise be spread across separate sources.

03Keep evidence and interpretation distinct

Support indicators, resident quotations and contextual information serve different purposes. A readable report needs to distinguish the underlying evidence from the conclusions drawn from it.

04Design around the next decision

The report ends with what the findings mean and suggested next steps, connecting the information to practical engagement planning.

04 / THE FULL PICTURE

Evidence &
limitations.

What this shows

The report structure described here is based on an example output supplied by the author. It demonstrates the breadth and organisation of the product, not independently verified predictive performance. The interactive sources below use invented data.

What to keep in mind

All places, figures and source descriptions in the interactive walkthrough are fictional. The demo is a new portfolio interface, not the original product. It does not reproduce the original report, employer datasets or private records, and its figures are not findings about real places.

ANOTHER SIDE OF MY WORK

Mapping dark matter

Machine learning · MSc research