Wearable EEG solutions allow portability and realtime measurements in uncontrolled conditions. However, for accurate interpretation of the EEG data, it is essential to determine artefact-contaminated data in EEG measurements so that the output is reliable and reproducible. In this paper, two data quality indicator approaches are proposed. The first approach, a pragmatic model, is based on statistical features and data-driven thresholding. The second approach is based on the same set of statistical features, but uses a regression model to predict data quality. The results were validated against EEG data recorded during uncontrolled laboratory and free-living conditions, manually annotated by experts. Both approaches show high accuracy in detecting artefactual data. Training of the two approaches on free-living conditions data showed better generalization when tested on different types of databases, i.e. uncontrolled laboratory and free-living. Although the accuracy in determining artefact-contaminated data is highest when using a window size of 8 seconds, the accuracy drop is minor when using shorter window size, facilitating realtime applications. Data quality indicators using relatively simple statistical features can distinguish with at least 80% accuracy between good and bad data. The two models perform equally across different databases. Complexity of both models allows for implementation in real-time, which is demonstrated using wearable EEG headset system available at IMEC.