The first level of the course is for people just starting to work with data — or who have counted things by eye their whole life. In eight short lessons you build a foundation: what a table is and why files break, how to find what you need, roll numbers into totals, and tell an honest number from a misleading one. No formulas — just a skill that stays with you in any work with numbers.
Every lesson is a template with demo data: ten minutes to complete, one skill that stays. Open it in the Templates section of Analytics Agent and press “Use this template” to take the lesson on your own data.
What Data Is: Rows and Columns
Shows: 30 coffee-shop purchases laid out in a table: a row is one event, a column is one property, and every column has its own type. After this lesson, any export turns into a picture you can read.
You get: the first step from “numbers floating around” to data you can count — the foundation for every lesson that follows.
Your First Table: Loading a File
Shows: loading a file and checking its “calling card”: how many rows, what period, what total. Three typical breakages show how a blank row, a foreign date format, and “money stored as text” quietly change the answer.
You get: the habit of checking incoming data in the first minutes — instead of catching distorted totals a week later.
Sorting and Filtering
Shows: sorting answers “bigger or smaller”, a filter answers “only these” — and neither changes the data, only your view of it: the top 10, the worst five, one category out of two hundred receipts — in a couple of moves.
You get: any ranking and any “how much X do we have” no longer requires asking someone else to count.
Your First Pivot Table
Shows: the four zones of any pivot — rows, columns, values, filters — and a totals matrix that checks itself: 524 rows compress into 12 sums with nothing lost.
You get: a skill that transfers to any tool, from spreadsheets to Analytics Agent — manual summaries are no longer needed.
Clean Data: Duplicates, Blanks and Typos
Shows: on a dirty file you see that duplicates and typos cost 7.8% of the money; the lesson shows how to remove duplicates, flag blanks, and unify spellings, keeping a log of every fix.
You get: you stop trusting numbers from a dirty export, you know how to clean it up — and you never count the same money twice.
Sum, Average, Minimum, Maximum
Shows: five simple counters — sum, count, average, minimum, maximum — and their trap: an average without “out of how many” misleads. A manager example shows why the best by deals can be the worst by average check.
You get: every “on average” now reads honestly to you — with an explicit base, not a magic number.
Average or Median: Why the “Average Salary” Lies
Shows: the difference between the mean (“all the money ÷ people”) and the median (“the person in the middle”): one outlier swings the mean 2.8× while the median holds still. The rule: money, time on site, waiting times are skewed data — they need the median.
You get: you no longer fall for the “average salary on the market” and can honestly assess your own business.
Shares and Percentages
Shows: “share” (a part of a whole) vs “growth in percent” vs “percentage points” — plus the tiny-base trap, where “it doubled!” is really 2.7 points of revenue. It also shows the quiet decline that one grand total hides.
You get: you always name the base of every percentage — what it was counted from and in which parts — so loud percentages stop being taken at face value.