logo

Data analytics & streaming

Real-time data pipelines and analytics, built on open source

We stream data from Odoo, your databases and your apps through Kafka as it changes. ClickHouse stores and models it, and your team sees the numbers on one dashboard, without per-seat licences.

Plan my data platform Ways to start

We reply within one business day.

Seconds

from source change to dashboard

Fixed price

scope agreed in writing before work starts

0

per-user BI licence fees

Change data capture → Kafka → ClickHouse → dashboards

Change data capture → Kafka → ClickHouse → dashboards

The open-source stack we build with

Apache Kafka

ClickHouse

PostgreSQL

Apache Airflow

dbt

Apache Spark

Apache Superset

Metabase

Grafana

MongoDB

Odoo

In short

We design and run open-source data platforms: change data capture from PostgreSQL, MySQL or Odoo into Apache Kafka, transformations with dbt and Airflow, fast analytics in ClickHouse, and dashboards in Apache Superset, Metabase or Grafana. Work starts with a fixed-price data platform assessment, then a pilot pipeline for one real use case before anything is scaled.

Services

What we build

One team for the whole path, from source systems to the dashboard on the wall.

Kafka streaming

Kafka clusters (self-managed, Strimzi on Kubernetes or managed), topics, schemas and Kafka Connect, sized and secured for production.

Change data capture

Database changes captured as events (Debezium) so analytics stay in sync without heavy nightly exports.

Pipelines and orchestration

Batch and streaming jobs in Airflow, with retries, alerts and lineage you can follow.

Data modelling

dbt models with tests and documentation, so every metric has one definition everyone trusts.

Analytics databases

ClickHouse for fast analytics on billions of rows; PostgreSQL or TimescaleDB where they fit better.

Dashboards and BI

Superset, Metabase or Grafana dashboards for management, sales, operations and finance, with access control.

Proof

Related work we delivered

Real projects. We name clients only with their permission.

Document AIArchive and records organisation (name withheld)

15 million archive images turned into searchable data

Challenge

Millions of scanned historical documents, partly handwritten, that could only be browsed image by image.

What we did

An extraction pipeline combining OCR, handwriting recognition and language models, with quality checks and a search index on top.

Result

15M images processed into structured, searchable data.

OCRHandwriting recognitionLLMsSearch
AI engineering at zd-inc.net →

Examples

Typical data analytics projects

OdooCDCClickHouseSuperset

Live sales and stock dashboard for Odoo

Management sees revenue, margin and stock by channel in near real time, without slowing down Odoo with heavy reports.

App eventsKafkaClickHouseGrafana

Event analytics for a SaaS product

Product and usage events streamed and queried in seconds, at a fraction of the cost of per-event analytics tools.

PostgreSQLDebeziumKafkaWarehouse

Replace nightly exports

Fragile overnight scripts replaced with a change stream, so reports are current and failures are alerted.

ERPCRMAirflow + dbtMetabase

One finance view across systems

Revenue, costs and cash from several systems modelled once, with tested definitions the finance team signs off.

How to start

How to start with data analytics

Infrastructure is billed by your cloud or hosting provider. Managed operations can run on your own servers.

Data platform assessment

Fixed priceagreed after a short call
  • About 2 weeks

  • Sources, volumes and use cases

  • Target architecture and costs

  • Pilot plan

Book the assessment
MOST POPULAR

Pilot pipeline

Fixed priceper phase, scoped in writing
  • One real use case end to end

  • CDC or batch ingestion

  • Models and one dashboard

  • Runbook and handover

Plan a pilot

Managed data platform

Monthly plansized to your needs
  • Monitoring and alerting

  • Upgrades and backups

  • New sources and models

  • Response times in writing

Ask about managed

Process

Getting started with data analytics

1

Assess

We map your sources, data volumes, the questions people need answered and what you spend today.

2

Pilot

One use case built end to end on the real data, so you see value and cost before scaling.

3

Scale and run

More sources and dashboards added in phases; we monitor, upgrade and support the platform.

Why it matters

Why data projects stall

Most companies have the data. The problem is getting it out of operational systems, trusting the numbers and keeping it fast and affordable.

Reports slow down production

Heavy queries on the ERP or app database slow everyone down. Analytics belongs on its own system.

Nightly exports break

Scripts that copy data overnight fail silently and leave dashboards stale.

Different numbers in every report

Without shared definitions, sales and finance argue about whose revenue is right.

Per-seat BI costs

Licences per viewer make it expensive to share dashboards widely.

Real time when it matters

Stock, fraud or operations dashboards need data within seconds.

Nobody owns the pipeline

Without monitoring and owners, data goes wrong and nobody notices.

When to keep it simple

An honest note before you book.

  • Small data volumes: built-in reports or a spreadsheet may be enough.

  • Daily reporting only: scheduled batch loads are simpler than streaming.

  • An established data team and warehouse: we can still help with specific pipelines.

Approaches

Reports on the production database vs a proper analytics stack

Area
Reports on production
Open-source analytics stack
Impact on users
Reports on production: Heavy reports slow the ERP or app.
Open-source analytics stack: Analytics runs on a separate, fast system.
Freshness
Reports on production: Either live and slow, or exported and stale.
Open-source analytics stack: Near real time through change data capture.
Trust
Reports on production: Each report calculates metrics its own way.
Open-source analytics stack: One tested definition per metric in dbt.
Cost to share
Reports on production: BI licences per viewer.
Open-source analytics stack: Open-source BI without per-seat fees.

FAQ

Data analytics FAQ

Straight answers. Ask us anything else on a call.

Do we need Kafka?

Not always. If hourly or daily data is fine, scheduled batch loads are simpler and cheaper. Kafka and change data capture pay off when you need data within seconds or many systems consume the same events.

Why ClickHouse instead of a cloud data warehouse?

ClickHouse is open source and very fast for analytics, and you can run it on your own servers or a managed service, so costs stay predictable. For some teams a cloud warehouse is the better fit; we compare both in the assessment.

Can we keep our current BI tool?

Yes. ClickHouse and PostgreSQL work with most BI tools. We only suggest Superset or Metabase if licence costs or limits are a problem.

Will this slow down our production database or Odoo?

No. Change data capture reads the database log rather than running heavy queries, and analytics run on a separate system.

Related

Related services

Odoo services

Next step

Talk to an engineer about data analytics

Send a few details, and pick up to 3 times if you would like a call. We reply within one business day.

You talk to the expert who would do the work

Written scope and price after the call

No obligation

info@opensource.consulting
+1 514 577 8599

OpenSource Consulting

Open source solutions integrator: we design, build, integrate and support Odoo, n8n, Zabbix, Grafana, Kubernetes and databases.

info@opensource.consulting

+1 514 577 8599

Canada · Netherlands · UAE

Services

© 2026 OpenSource Consulting. All rights reserved. · Privacy · Cookies · Terms

Our brands: elasticsearch.expert · zd-inc.net · opsfinitive.io

Plan my data platform