AI data analysis for your CSV and Excel files
Upload a dataset, ask the AI assistant what is worth testing, and run the analysis yourself — descriptive statistics, correlation, regression, t-tests, ANOVA and chi-square in one workspace.
Statistical analysis you can run in Theta
Each method answers a different kind of question. Here is what you ask and what you get back.
Descriptive statistics
What does each column look like?
Count, mean, median, standard deviation, minimum, maximum and quartiles for every numeric column.
Correlation matrix
Which numeric columns move together?
A Pearson correlation matrix across your numeric columns, which you can also view as a heatmap.
Linear regression
How well does one variable predict another?
Slope, intercept, R² and the number of observations for a simple regression of Y on X.
t-test (two groups)
Do two groups differ on a numeric measure?
The t statistic, with the mean and sample size of each group.
ANOVA (three or more groups)
Do several groups differ on a numeric measure?
A one-way ANOVA with the F statistic, degrees of freedom and the mean of each group.
Chi-square
Are two categorical columns related?
The chi-square statistic, degrees of freedom and the observed contingency table.
How the AI assistant helps
- Suggests which test or chart fits your question, and which columns to use.
- Explains your dataset and your results in plain language, followed by a more formal interpretation.
- Points out patterns and anomalies worth a closer look.
- Works from your dataset's structure rather than giving generic advice.
The assistant recommends; you stay in control. You choose the variables and run the test in the Analysis panel.
Good to know
Theta reports test statistics (t, F, χ²), degrees of freedom and group means, and flags whether a result looks significant. It does not report exact p-values, so confirm important results in dedicated statistical software before you publish or make high-stakes decisions.
That makes Theta well suited to exploring data, learning statistics and getting a quick first read.
From file to result in five steps
- 01
Upload your file
Drop in a CSV, TXT, XLSX or XLS file. Theta reads the columns and detects whether each one is numeric, categorical or a date.
- 02
Look at your data
Browse and search the rows in the Dataset view so you know what you are working with before you test anything.
- 03
Ask the AI assistant
Describe your question and the assistant suggests a test or chart and which columns to use.
- 04
Run the analysis
Pick the test and the variables in the Analysis panel and run it. The calculations happen in the app.
- 05
Export the result
Download any result as PDF, CSV, TXT or JSON, or generate a PDF report of descriptive statistics and correlations.
When this is useful
- Checking coursework or survey data before you write it up
- Comparing two or more groups on a measure, such as scores by class or sales by region
- Looking for relationships between numeric columns in a business spreadsheet
- Getting a first, structured look at a dataset you have just received
AI data analysis questions
What is AI data analysis?
It means using an AI assistant to help you decide what to analyze and to explain what you find. In Theta the assistant recommends tests and charts and explains datasets and results in plain language, while the statistics themselves are calculated in the app.
What file types can I analyze?
You can upload CSV, TXT, XLSX and XLS files.
Do I need to write code?
No. You choose columns and tests from menus. There is nothing to script or install.
Which statistical tests does Theta support?
Descriptive statistics, a correlation matrix, simple linear regression, a two-group t-test, one-way ANOVA and a chi-square test.
Can I export my results?
Yes. Each analysis result can be exported as PDF, CSV, TXT or JSON, and the Report view creates a PDF with methodology, descriptive statistics, correlations and an interpretation section.
What does the AI assistant see?
It receives your dataset's name, row count, column names and types, and a small sample of rows (the first five) so it can give relevant suggestions. Avoid uploading data you are not allowed to share.
Try it on your own data
Upload a file and ask the assistant where to start, or load the sample dataset to look around.
Also see online data visualization, the Learning Center and the FAQ.