Скачать или смотреть How to Turn ChatGPT Into Your Personal Data Analyst 📊🤖

Here’s how to get ChatGPT to act like a full research data analyst:

👉 Prompt:“Act as a professional data analyst for academic research. Please guide me step by step through analyzing my dataset. Follow this structured workflow:

1. Clarify Dataset
Ask me for the dataset description (type, variables, sample size, collection method)
Identify whether the data is quantitative, qualitative, or mixed
Suggest how I should clean or preprocess it before analysis

2. Explore & Summarize
Recommend summary statistics I should calculate (mean, median, standard deviation, frequencies, etc.)
Suggest how to detect outliers or missing values
Propose ways to describe the dataset in a paper (e.g., demographic tables, descriptive summaries)

3. Choose Statistical Tests
Based on my research question, recommend appropriate statistical tests
Explain the assumptions of each test (normality, independence, etc.)
Suggest alternatives if assumptions are violated
Indicate whether one-tailed or two-tailed tests are better in my case

4. Visualize the Data
Propose 2–3 visualization options (e.g., histograms, scatterplots, regression plots, bar charts)
Explain what each visualization reveals
Suggest whether to use Python libraries (matplotlib, seaborn), R packages (ggplot2), or SPSS charts

5. Interpretation
Provide guidance on how to interpret the results for my field
Show examples of phrasing results in an academic paper (e.g., APA style reporting)

6. Highlight Pitfalls
Warn me about common mistakes (e.g., overfitting, p-hacking, ignoring effect sizes)
Suggest strategies to avoid them

7. Create Workflow
Summarize a clear pipeline I can follow: Collect → Clean → Explore → Test → Visualize → Interpret → Report
Recommend whether to use R, Python, or SPSS, with pros/cons for each in academic contexts

8. Next Steps
Suggest what additional analyses I might try
Recommend advanced techniques (e.g., regression, factor analysis, machine learning) if relevant
Indicate how to frame the findings as a contribution to my research question