How to Analyze a CSV File (Without Being a Data Analyst)
How to analyze a CSV file quickly: opening it correctly, fixing encoding and delimiter problems, cleaning columns, profiling the data, charts that answer real questions, and tools that do it in your browser.

On this page
Your till, bank, online shop or accounting system can probably export data as a CSV file. It's the universal format: plain text, one row per line, values separated by commas. The trouble is what happens next. Double-click it and the dates are scrambled, the accents turn into symbols, or everything sits in one column. This guide shows how to analyze a CSV file properly: open it the right way, clean it, understand what's inside, and answer real questions.
What a CSV file is#
A CSV file stores a table as plain text: each line is a row, and values in a row are separated by a delimiter, usually a comma. The first line normally holds the column names. The common format is described in RFC 4180, though real exports vary a lot.
date,product,units,unit_price,revenue,branch
2026-10-01,Coffee 250g,4,5000,20000,Toronto
2026-10-01,Honey 500g,2,7000,14000,Ottawa
Step 1: Open it correctly#
Most problems start with double-clicking. Instead, import the file, so you can choose how it's read:
- Delimiter: comma, semicolon or tab. European systems often use semicolons, because commas are decimal separators there.
- Encoding: choose UTF-8, so names like "Muhoza" stay fine and "Hélène" doesn't become "Hélène".
- Column types: keep phone numbers and IDs as text, so leading zeros and long numbers aren't lost or turned into scientific notation.
In Excel, use Data, then From Text/CSV. In Google Sheets, use File, Import. Browser tools usually detect the delimiter for you.
Step 2: Check what you've got#
Before any analysis, get a feel for the file:
- How many rows and columns?
- What does each column contain, and is it the type you expect: dates, numbers, text?
- What period does it cover? Check the earliest and latest dates.
- Are there blank cells, and in which columns?
- Are there obvious oddities, such as negative quantities or prices of zero?
This is called profiling the data. Doing it by hand takes a while; a profiling tool does it instantly. The Tekservix AI data analysis tool reads CSV, TSV or JSON files up to 10 MB in your browser, profiles every column, charts it and adds AI findings, sending only a summary for the AI part. If your data is in Excel, save it as CSV first.
Step 3: Clean the data#
Most real exports need some tidying:
- Remove extra header or footer lines, such as report titles or totals at the bottom, which break analysis.
- Standardise names: "Toronto", "TORONTO" and "Tor" should be one value.
- Fix dates into one format, such as 2026-10-01.
- Remove currency symbols and text from numbers, so "C$20,000" becomes 20000.
- Handle blanks: decide whether a blank means zero, unknown or an error.
- Remove duplicates, which often appear when two exports overlap.
Keep the original file untouched and clean a copy, so you can always start again. For messy text patterns such as phone numbers, our regex tester guide shows how to check them in bulk.
Step 4: Ask specific questions#
Analysis goes faster with a clear question. For a sales export:
- Which products bring the most revenue, and which the least?
- How do sales change by day of the week, or by month?
- Which branch is growing fastest?
- What's the average order value, and is it rising?
Pick one or two, answer them, then move on. Our guide to analysing data with AI has more questions by type of business, and how to turn findings into actions.
Step 5: Summarise and chart#
Most answers come from a few simple operations:
- Totals and averages by a category: revenue by product, orders by branch.
- Counts: how many orders, customers or items.
- Grouping by time: by day, week or month.
- Sorting: top ten and bottom ten.
In a spreadsheet, a pivot table does most of this. Then pick the chart that fits the question: a line chart for trends over time, a bar chart for comparing categories, and avoid pie charts with more than a few slices.
Step 6: Check before you trust#
- Does the total match another source, such as your accounts or the system's own report?
- Are refunds or cancelled orders included when they shouldn't be?
- Is a strange spike real, or a duplicate import?
- Is the period complete, or does the export stop mid-month?
When CSV isn't the right format#
CSV is ideal for flat tables. Data with nested structure, such as orders that each contain several items with options, is often exported as JSON instead. Our JSON formatter guide explains how to read it, and many analysis tools accept a JSON array of records directly.
Privacy#
Exports often contain customer names, phone numbers and addresses. Remove columns you don't need before analysing or sharing, and delete working copies when you're done. Prefer tools that process the file on your own device.
A worked example#
Imagine a shop owner exporting ten months of orders from an online store. Double-clicking shows everything in one column, so they import it choosing a semicolon delimiter and UTF-8. The profile shows 4,812 rows, dates from January to October, and blanks in the "discount" column, which they decide mean no discount. They remove 40 duplicate rows from an overlapping export and standardise branch names. A monthly revenue chart shows a clear dip in April; the owner remembers a two-week stock-out of the best-selling coffee, which the product breakdown confirms. They raise that product's reorder level, and the question "why was April slow?" finally has an answer.
Frequently asked questions
What is a CSV file?
A CSV (comma-separated values) file is a plain text file where each line is a row and values are separated by commas or another delimiter. Most systems, such as tills, banks, shops and accounting tools, can export data as CSV.
Why does my CSV look wrong when I open it?
Usually because of the delimiter or encoding. Some systems use semicolons instead of commas, and accented characters break when the file isn't opened as UTF-8. Import the file and choose the delimiter and encoding instead of double-clicking it.
How can I analyze a CSV file without Excel?
Use a browser-based CSV analysis tool, Google Sheets, or a free spreadsheet program. Tools that profile the file automatically show each column's type, totals, ranges and missing values without any formulas.



