EuroData Score performs well-defined calculations on the data you provide. The results are only as reliable as that data and the way they are interpreted.
Key limitations
- Results describe your data only. Statistics summarize the records in the dataset. They do not automatically apply to a wider population unless the data is a representative sample.
- Historical patterns do not predict the future. Trends, frequencies and recurring values show what happened in the past. EuroData Score does not forecast outcomes and no result should be read as a guarantee of future values or events.
- Correlation is not causation. A strong correlation means two variables moved together in this data; it does not show that one causes the other.
- Unusual values are not necessarily errors. Values outside Tukey's fences or with a large z-score are flagged for review, not removed.
- Data quality matters. Missing, invalid and duplicate records change counts, sums and averages. Review the data quality panel before drawing conclusions.
- Small samples are unstable. With few records, statistics such as standard deviation, skewness or correlation can change substantially when a single value changes.
- Automated observations are rule-based. The observations on the Insights page come from fixed, documented rules (for example "|skewness| ≥ 1"). Each one states its rule so you can verify it.
Not professional advice
EuroData Score provides statistical analysis and informational insights. It is not financial, legal, medical or other professional advice. Consult a qualified professional before making important decisions based on data.