Names your colleagues recognise
Rename cryptic columns, set the type so a date is treated as a date and an amount as a number, and hide the fields nobody in the business ever asks about.
Data preparation
Settle the data once, not in every chart.
Join tables, add the figures no system delivers, straighten out field names – in the same place where the dashboards are built. Whatever is clear in the model, no chart has to work around later.
30 days free · No credit card required
The order of work
From several tables to one model
-
Step 1: Bring the tables together
Start in the model, not on a canvas. Pick the connected sources you need – the order table out of a database, a cost list from a file, the records a cloud service keeps – and put them side by side.
-
Step 2: Say how they belong to each other
Customer number, article number, order date: point at the field two tables have in common and set the relationship in the interface. Nothing is exported and nothing is copied into a second system for this.
-
Step 3: Calculate what nobody delivers
Margin, days outstanding, revenue per head: a calculated field is defined once in the model and then behaves like any other field. It can be filtered, grouped and dropped onto a canvas.
-
Step 4: Give the model a name
The result is the field list your colleagues build from. Joins, calculations and tidy names come with it, and the scheduled refresh reads the sources again without anyone touching the model.
Working on a source
What you can change before a single number is shown
Rows that should not count
Filter at the source: test entries, cancelled orders, the years before the system changed. They stay out of every chart instead of being excused in a footnote.
Prepared once, applied every time
The steps you define are part of the model. When the scheduled refresh reads the source again, the joins, filters and calculations are applied to the new rows as well.
One change, not twenty
Correct a calculated field in the model and the widgets that use it follow. A definition of margin lives in one place, not in each department’s copy.
Asked before the first model
Four things worth clearing up
Do we need a data warehouse before we can start?
No. The sources are read where they are and combined in one model. If a warehouse already exists in your company, it is simply one more source among the others.
Is a copy of our data created somewhere?
No export and no copying. smartlife BI reads the sources directly, and the preparation describes how they are combined – not a second store of the same rows.
May we write SQL where we prefer it?
Yes. The preparation step is where SQL has its place, for the query someone on your team would rather write than click. Everything else is done visually, and no model requires a line of it.
Who does this – IT or the department?
Both work, and that is the point. Joining tables and adding a calculated field needs no code, so the person who knows what the numbers mean can do it without a ticket.
Bring your two most awkward tables
In a 30-minute session we join them, add the field you are missing and show you the model behind the chart.