Visveswaran Data Lab
Visveswaran /data
visveswaran.in/data · data analyst · data domain consultant

Data that breaks heals itself, and answers people act on.

I clean, connect and explain data: patient and trial records, master data across systems, and the business questions sitting on top of them. Every project here runs on real public data, with its raw files, source code and a written account of how I built it.

Explore the projectsRun the self-healing demo
healing_engine · pass 1InvalidMissingDuplicateHealed
Cells scanned0
Issues found0
Auto-healed0
Selected work

Seventeen projects on real data

Life-sciences studies, data cleansing and de-duplication, business decisions and visual analytics. Each one ships with a project document, a data sheet and the raw data it was built from.

Approach

How I work with data

The same five stages run through every project on this page, from a one-off analysis to an autonomous data quality platform.

  1. 01

    Profile

    Measure completeness, validity, consistency and uniqueness before anyone touches a value.

  2. 02

    Diagnose

    Trace each defect to its root: a missing rule, a unit mix-up, a site-specific habit, a broken feed.

  3. 03

    Heal

    Fix with rules mined from the data itself. Confident fixes run automatically; the rest go to a data steward.

  4. 04

    Govern

    Every change gets an audit trail, an owner and a confidence score, so the fix is trusted and repeatable.

  5. 05

    Explain

    Turn clean data into views and recommendations people read in seconds and act on.

Domains

Where I can help

Life-sciences & clinical data

Patient record matching, multi-site data harmonisation, trial reporting tables, survival analysis and surveillance data cleaning.

Record linkageSurvival analysisTrial tables

Master data & data quality

Material, customer, vendor and site master design, SAP material master structures (MARA, MARC, MAKT, MARM), DQ scorecards and steward workflows.

SAP MMMatch & mergeGolden record

Business problem solving

Experiments, churn and retention economics, operational KPIs: a clear question, a defensible method and a recommendation with a way to measure it.

A/B testingChurn modelsKPI design

Analytics & visualization

Dashboards, 3D and interactive data stories in Python, SQL, D3 and Power BI, designed for the question the audience is asking.

PythonSQLD3.jsthree.js

Have data that needs fixing, or a question that needs answering?

I work with teams on data quality programmes, master data clean-ups, clinical and life-sciences data, and analytics that people actually use.

Get in touchSee a life-sciences project