Data jargon in plain English: 20 terms every career changer should know
Data job adverts and course pages are full of words that sound more complicated than they are. Most of them describe simple ideas you may already use at work just without the label.
So we’ve compiled 20 terms you’ll meet early on and organised them into categories so you can see how they build on from each other: what data looks like, how you get it ready, how you ask it questions and how you show the answer.
What data looks like
- Dataset: a collection of related data put together for a purpose, often spread across several tables. A year of sales, customers and products is a dataset.
- Table: data arranged in rows and columns, like a single sheet in a spreadsheet.
- Record: everything stored about one thing, such as one customer or one order. In a table, a record is a row.
- Field: one piece of information stored for every record, such as an email address. In a table, a field is a column.
- Primary key: a field that identifies each record uniquely, so no two records share it. A customer ID is a common example.
- Data type: the kind of value a field holds, such as a number, text or a date. It decides what you can do with the field: you can add up numbers, but not names.
- CSV: a plain text file that separates values with commas. It’s a common way to move data out of one system and into another.
Getting data ready
- Data cleaning: fixing mistakes, gaps and inconsistencies so the data can be trusted. Think dates written three different ways, the same customer spelt two ways and blank rows.
- Data wrangling: reshaping raw data into the form your analysis needs. Cleaning is one part of it.
- Duplicate: a record that appears more than once. Duplicates make totals look bigger than they are, so they’re usually removed during cleaning.
- Missing values: empty cells where you expected an entry. You decide whether to fill them, flag them or leave the record out, and you note what you did.
- Outlier: a value far away from the rest, such as one £9,000 order in a list of £20 orders. It could be a typo or the most interesting thing in the data, so check before you act.
Asking questions of your data
- Query: a request for specific data from a database. Analysts usually write queries in SQL, which reads close to plain English.
- Filter: showing only the rows that match a condition, such as orders over £100.
- Join: combining two tables by matching a field they share. Link an orders table to a customers table using the customer ID, and you can see who bought what.
- Lookup: an Excel function that finds a value in one table and brings back related information. VLOOKUP is the older version; XLOOKUP is newer and more flexible.
- Aggregation: rolling many values up into one figure, such as a total, a count or an average. “Sales by region” is an aggregation.
- Pivot table: an Excel tool that summarises a large table in seconds, for example total sales by month and region.
Showing the answer
- Dashboard: one screen of charts and key numbers, often updating automatically, so people can see how things are going at a glance.
- KPI: key performance indicator. A number that shows progress towards a goal, such as monthly sales against target.
How to use this list
Next time you read a data analyst job advert, highlight every term you recognise from this list. The ones you can’t explain yet make a ready-made learning plan.
To put the words into practice, download a free dataset from data.gov.uk or the Office for National Statistics. Try a filter, an aggregation and one chart.
If a dataset includes information about people, remember that personal data is protected by law. The Information Commissioner’s Office explains what counts as personal information under UK GDPR.
Next step: Want all 68 terms in one place? Get our free Data Jargon Buster, or book a free 15-minute chat with a career advisor to talk about a move into data.
