Data analytics: Tableau

Sefism.

Chapter 02

Spreadsheets properly

Excel or Google Sheets used like an analyst: structured tables, formulas you can audit, lookups, pivot tables and cleaning that someone else can follow.

Relative and absolute references

Also called $A$1, mixed reference, cell reference, relative reference, absolute reference

What
A relative reference (A2) shifts when you copy a formula; an absolute one ($A$2) stays fixed; a mixed one ($A2 or A$2) fixes only the column or row.
Why
Copying formulas down thousands of rows is how spreadsheets scale. Wrong anchoring silently produces wrong numbers.
EXCEL
=B2*$F$1        tax rate fixed in F1
=B2/SUM(B$2:B$100)   share of column total

Watch out

Forgetting the $ on a rate or lookup range, so rows further down point at empty cells.

Structured tables

Also called Excel table, Ctrl+T, structured references

What
A range turned into a named table, whose formulas refer to columns by name and which grows automatically when you add rows.
Why
Formulas, pivot tables and charts built on a table pick up new data without editing ranges, which removes a classic source of stale reports.
EXCEL
=SUM(Sales[Amount])
=[@Quantity]*[@UnitPrice]

Watch out

Referencing A2:A500 by hand, then adding row 501 that no total includes.

SUMIFS and COUNTIFS

Also called AVERAGEIFS, conditional sum

What
Functions that sum, count or average only the rows that meet one or more conditions.
Why
They answer most quick business questions, such as sales by region for a month, without a pivot table.
EXCEL
=SUMIFS(Sales[Amount], Sales[Region], "North",
       Sales[Date], ">="&DATE(2026,1,1))
=COUNTIFS(Sales[Status], "Refunded")

Watch out

Text that looks like a match but has a trailing space, so the condition quietly skips it.

XLOOKUP

Also called VLOOKUP, INDEX MATCH, lookup

What
Looks up a value in one column and returns the matching value from another, with a fallback when nothing is found. It works in current Excel and Google Sheets.
Why
Bringing a column from one table into another (a customer region, a product price) is one of the most common tasks in analysis.
EXCEL
=XLOOKUP([@CustomerID], Customers[ID],
        Customers[Region], "Not found")

Watch out

Using VLOOKUP with approximate match by default, which returns a wrong row instead of an error.

Text and date cleaning functions

Also called TRIM, TEXT, DATEVALUE, SPLIT, PROPER, SUBSTITUTE

What
Functions that fix the usual mess: extra spaces, mixed case, text that should be numbers or dates, and codes that need splitting.
Why
Dirty text breaks lookups, counts and pivot tables. Cleaning in helper columns keeps the original data intact and the steps visible.
EXCEL
=TRIM(PROPER(A2))            clean a name
=DATEVALUE(B2)               text to date
=TEXT(C2, "yyyy-mm")         month label

Watch out

Overwriting the raw column with cleaned values, so nobody can check what changed.

FILTER, UNIQUE and SORT

Also called dynamic arrays, spill range

What
Functions that return a whole range of results that spill into neighbouring cells: a filtered list, the distinct values or a sorted copy.
Why
They replace many manual copy and paste steps with formulas that update when the data changes.
EXCEL
=UNIQUE(Sales[Region])
=SORT(FILTER(Sales, Sales[Amount]>1000), 3, -1)

Watch out

Typing into the spill area below the formula, which blocks it with a #SPILL! error.

Pivot tables

Also called PivotTable, pivot

What
A tool that summarises a table by dragging fields into rows, columns, values and filters, such as total sales by region and month.
Why
They are the fastest way to explore a new dataset and answer "how much, by what" questions, and they teach the group-by thinking SQL and BI tools use.
Example
Rows: product category. Columns: month. Values: sum of revenue and count of orders. Filter: country.

Watch out

Summing a column that is already an average or a percentage, which gives a meaningless total.

Data validation

Also called drop-down list, input rules

What
Rules on cells that limit what can be entered, such as a drop-down of allowed regions or dates within a range.
Why
Stopping bad data at entry is far cheaper than cleaning it later, especially in sheets that several people fill in.

Watch out

Adding validation after months of free typing and assuming the old rows now follow the rules.

Raw, working and output tabs

What
Keeping untouched source data on one tab, cleaning and calculations on another, and the summary or chart on a third, with a notes tab describing each.
Why
Anyone, including you in three months, can trace a number back to its source and rerun the steps with new data.

Watch out

Hard-coding numbers typed from another report into a formula, so nobody can see where they came from.

02 · Spreadsheets properly2 / 9