↓ Skip to main content
Lough on Data
Lough on Data

Lough on Data

Reliable data systems that reflect the business, built for the team that runs them
#

Lough on Data helps growing teams and early startups turn brittle pipelines into reliable foundations—so people can trust the data they use to decide, build, and operate.

When data arrives late, breaks silently, or needs constant manual repair, the problem is rarely one bad job. It is the system around it: how data enters, how it is modeled, how failures are found, and who can operate it when something changes.

Tell me what’s brittle →

Build in public

View GitHub profile →
GitHub contributions over the last 12 months

Reliability is a system

Build foundations around the work.

When data arrives late, breaks silently, or needs constant manual repair, the problem is rarely one bad job. It is the system around it.

  1. SourcesWhere data begins
  2. IngestionHow it arrives
  3. ModelsHow it becomes useful
  4. Quality & operationsHow it stays dependable
  5. Trusted decisionsWhat the system enables

Focused work

Where I can help.

  1. 01

    Platform Assessment

    Find the architecture, reliability, cost, modeling, and delivery risks worth addressing first.

  2. 02

    Pipeline Buildout

    Build dependable ingestion, transformation, orchestration, and observability paths.

  3. 03

    Warehouse & Modeling

    Make analytical storage, models, metrics, and semantic layers easier to trust and use.

  4. 04

    Data Reliability

    Add the quality controls, documentation, and operating practices that keep systems useful.

Explore services

Public architecture evidence

Databox, in the open.

A single-operator data platform that brings public bird observations, weather, and streamflow into one queryable warehouse.

Dagster global asset lineage showing ingestion assets feeding analytics and environmental observations before birding_agent
Dagster asset lineage makes the path from ingestion to derived observations and birding analysis inspectable.
  1. eBird, NOAA, and USGS
  2. Shared spatial grain
  3. DuckDB warehouse
  • Public sources join at H3 cells by day.
  • Quality checks can gate downstream materialization.
  • Models, columns, checks, and lineage stay inspectable.
Inspect Databox

The operator behind it
#

I’m Connor Lough, an independent data engineer working remotely with teams worldwide. I write code to solve useful problems—the best result is usually a system that makes someone else’s work quieter and clearer.

More about how I work →

Start with the problem
#

Email me a little context: what is unreliable, who relies on it, and what outcome you need. We’ll talk through the fit and, if it makes sense, define a focused next step.

Tell me what’s brittle →