Introducing DataLAB: SQL-Based Analytics from Snaplytics

The problem we set out to solve

Across South Africa and the broader continent, strong analysts still spend too much of their day inside spreadsheets. They know SQL. They understand their data. But they often do not have access to practical software that lets them move from data access to transformations, pipelines, and predictive workflows in one place.

We built DataLAB to change that.

What is DataLAB?

DataLAB is Snaplytics' desktop-first analytics product built around SnapQL, our SQL-based language for analytical workflows. If you can write a SELECT statement, you can:

• Query datasets across files and connected sources
• Build transformations and reusable pipelines
• Train and score machine learning models
• Validate and match records across datasets
• Track experiments and compare analytical runs

All from a consistent query surface.

A quick tour

Analytics

Imagine an operations analyst or commercial analyst loading a monthly sales extract and needing a quick regional view before a review meeting. The point is to move from raw data to a decision-ready summary without exporting into another tool first.

SELECT region, SUM(revenue) AS total,
       COUNT(DISTINCT customer_id) AS customers
FROM sales
GROUP BY region
ORDER BY total DESC;

DataLAB gives teams a SQL-first workflow with support for CTEs, window functions, PIVOT, and the broader SnapQL command surface.

Machine learning

Now think about the same team handing a churn or risk problem to an analyst who understands the business but does not want to build a Python project just to test a model idea. The workflow stays in the same working surface:

CREATE MODEL churn_predictor
    USING RandomForest
    ON customer_data
    PREDICT churned
    FEATURES tenure, monthly_charges, support_calls;

Then score a fresh customer extract or a new operational snapshot:

PREDICT USING MODEL churn_predictor
    ON new_customers
    AS churn_predictions;

Pipelines

This is where teams usually feel the pain most clearly. A month-end report, weekly export, or recurring management pack starts as one useful query and quickly turns into a brittle manual process. Pipelines give that work a reusable home:

PIPELINE monthly_report(@month DEFAULT '2026-03'):
    LOAD "sales.csv" AS sales WITH detect_types=true

    WITH department_totals AS
        SELECT department, SUM(amount) AS total
        FROM sales
        WHERE month = @month
        GROUP BY department

    EXPORT department_totals TO BROWSER AS monthly_report_output
END PIPELINE

Run the saved workflow when you need it:

CALL monthly_report('2026-03');

Validation and matching

The value becomes even more practical when two datasets have to agree. A team can move from imported datasets to a structured matching workflow without rebuilding the logic in spreadsheets:

MATCH order_lines TO shipment_lines
    ON order_id = order_id
    COMPARE quantity
    TOLERANCE 0.01;

Built for practical rollout

DataLAB is deliberately desktop-first, giving analysts a controlled workspace close to source data and local compute. Its server-ready architecture extends the same governed workflows to browser and team environments when broader access is needed.

See DataLAB in action

Analytics teams, operations teams, and data teams can use DataLAB to connect SQL-based analysis with transformation, pipelines, machine learning, and governed review. Watch Demo or Book Demo around your workflow.