← Selected work / 03

Workflow automation · Procurement analytics

Digital Purchase Request Tracking Portal

A clearer path from purchase request to purchase order — with operational insight at every step.

LIVE · INTERACTIVE APPLICATION

LIVE APPLICATION

Follow the workflow. Explore the analytics.

Interactive portfolio demo using fully synthetic data.

A 30-SECOND TOUR
  1. Create and assign a fictional purchase request.
  2. Switch demo persona, record progress and link a synthetic PO.
  3. Filter the analytics and download a matching Excel report.

01 / PROBLEM

Make procurement progress visible.

Manual purchase-request follow-up creates fragmented visibility across request creation, assignment, processing and PO conversion. Teams need a clear view of ownership, status and the requests that need attention.

02 / SOLUTION

Connect operational work with analytics.

I built a full-stack Flask application that centralizes the PR lifecycle and combines structured workflow with operational analytics. This independently authored public demo lets visitors explore the process with fictional records and their own resettable session.

03 / WORKFLOW

From request to purchase order.

  1. 01PR Created
  2. 02Assigned
  3. 03Under Process
  4. 04Operational Updates
  5. 05PO Linked
  6. 06PO Issued
Assign a request, record progress and estimates, then link a synthetic purchase order.

04 / DEMO PERSONAS

Explore each side of the workflow.

01

Demo Admin

Review all requests, create a fictional request, assign a purchaser and follow progress across the register.

02

Demo Purchaser

Review assigned requests, record processing updates and preset notes, adjust estimates and link a synthetic PO.

These selectable personas demonstrate role-based behavior; they are not production authentication or real user accounts. Each visitor’s changes stay within their own demo session.

05 / ANALYTICS

Understand volume, progress and ownership.

PR lifecycle KPIs

Explore total and open PRs, PR-to-PO conversion and the proportion of requests linked to a PO.

PO status and expected dates

Review completed PO snapshots and open POs past their expected date. These describe fictional order status, not proof of receipt or delivery.

Operational distributions

Inspect status, branch, category and assignee distributions, the assignee-by-status matrix and monthly request trends.

All displayed counts and amounts describe synthetic scenarios, not business achievements.

06 / WORKFLOW INTELLIGENCE

Keep every update in context.

Assignment and controlled status transitions establish ownership and progress. Timeline history, preset notes and editable estimates keep the request’s context visible. Linking a PO captures the request lines and sets its status to PO Issued.

Filters and search update both the register and analytics. Sorting changes the order, while pagination keeps the register manageable. Reset returns the visitor to the deterministic baseline.

07 / REPORTING

A report that matches the view.

Download deterministic XLSX reports for the complete filtered scope across all pages. Three sheets cover the request dashboard, request items and PO summary, with estimates and PO totals aligned to the same records.

Identical logical state produces byte-identical workbooks. Reports are generated in memory and contain only synthetic data.

08 / ARCHITECTURE

A self-contained analytics application.

  1. 01Browser · HTML / CSS / JavaScript
  2. 02Render · Gunicorn
  3. 03Python · Flask
  4. 04Deterministic synthetic data
  5. 05Visitor-isolated session state
  6. 06Workflow · Analytics · XLSX
Flask reconstructs each visitor’s scenario from deterministic data and session actions. Gunicorn serves the application on Render; no database or persistent disk is required.

09 / PUBLIC DEMO

Portfolio Demo · Synthetic Data Only.

A fictional procurement environment

No production database. No real purchase requests. No real supplier or company information. No production integrations.

Try the workflow, explore the analytics and return here to inspect the implementation.

NEXT CASE STUDY

COGS Data Pipeline & Analytics ETL
↑