Internet technologies have demonstrated their value for the early detection and prediction of epidemics. In diverse cases, electronic surveillance systems can be created by obtaining and analyzing on-line data, complementing other existing monitoring resources. This paper reports the feasibility of building such a system with search engine and social network data. Concretely, this study aims at gathering evidence on which kind of data source leads to better results. Data have been acquired from the Internet by means of a system which gathered real-time data for 23 weeks. Data on influenza in Greece have been collected from Google and Twitter and they have been compared to influenza data from the official authority of Europe. The data were analyzed by using two models: the ARIMA model computed estimations based on weekly sums and a customized approximate model which uses daily sums. Results indicate that influenza was successfully monitored during the test period. Google data show a high Pearson correlation and a relatively low Mean Absolute Percentage Error (R = 0.933, MAPE = 21.358). Twitter results are slightly better (R = 0.943, MAPE = 18.742). The alternative model is slightly worse than the ARIMA(X) (R = 0.863, MAPE = 22.614), but with a higher mean deviation (abs. mean dev: 5.99% vs 4.74%).

Syndromic surveillance with the use of Internet data has been used to track and forecast epidemics for the last two decades, using different sources from social media to search engine records. More recently, studies have addressed how the World Wide Web could be used as a valuable source for analysing the reactions of the public to outbreaks and revealing emotions and sentiment impact from certain events, notably that of pandemics. The objective of this research is to evaluate the capability of Twitter messages (tweets) in estimating the sentiment impact of COVID-19 cases in Greece in real time as related to cases. 153,528 tweets were gathered from 18,730 Twitter users totalling 2,840,024 words for exactly one year and were examined towards two sentimental lexicons: one in English language translated into Greek (using the Vader library) and one in Greek. We then used the specific sentimental ranking included in these lexicons to track i) the positive and negative impact of COVID-19 and ii) six types of sentiments: Surprise, Disgust, Anger, Happiness, Fear and Sadness and iii) the correlations between real cases of COVID-19 and sentiments and correlations between sentiments and the volume of data. Surprise (25.32%) mainly and secondly Disgust (19.88%) were found to be the prevailing sentiments of COVID-19. The correlation coefficient (R²) for the Vader lexicon is −0.07454 related to cases and −0.,70668 to the tweets, while the other lexicon had 0.167387 and −0.93095 respectively, all measured at significance level of p < 0.01. Evidence shows that the sentiment does not correlate with the spread of COVID-19, possibly since the interest in COVID-19 declined after a certain time.

A multidisciplinary journal that focuses on public health and technology, public health informatics, mass media campaigns, surveillance, and innovation in public health practice and research. Also dedicated to rapid open data sharing during epidemics.

Background: In recent years new forms of syndromic surveillance that use data from the Internet have been proposed. These have been developed to assist the early prediction of epidemics in various cases and diseases. It has been found that these systems are accurate in monitoring and predicting outbreaks before these are observed in population and, therefore, they can be used as a complement to other methods. In this research, our aim is to examine a highly infectious disease, measles, as there is no extensive literature on forecasting measles using Internet data, Methods: This research has been conducted with official data on measles for 5 years (2013–2018) from the competent authority of the European Union (European Center of Disease and Prevention - ECDC) and data obtained from Google Trends by using scripts coded in Python. We compared regression models forecasting the development of measles in the five countries. Results: Results show that measles can be estimated and predicted through Google Trends in terms of time, volume and the overall spread. The combined results reveal a strong relationship of measles cases with the predicted cases (correlation coefficient R= 0.779 in two-tailed significance p< 0.01). The mean standard error was relatively low 45.2 (12.19%) for the combined results. However, major differences and deviations were observed for countries with a relatively low impact of measles, such as the United Kingdom and Spain. For these countries, alternative models were tested in an attempt to improve the results. Conclusions: The estimation of measles cases from Google Trends produces acceptable results and can help predict outbreaks in a robust and sound manner, at least 2 months in advance. Python scripts can be used individually or within the framework of an integrated Internet surveillance system for tracking epidemics as the one addressed here.

Very short-term forecasting of solar photovoltaic energy production at national scale is challenging due to the high variability and spatial aggregation of generation across large territories. This paper evaluates Temporal Convolutional Networks (TCN) — a deep learning architecture based on causal and dilated convolutions — for nowcasting national-level solar production at one-hour and fifteen-minute horizons, using data from Spain sourced from the European Network of Transmission System Operators for Electricity. Multivariate models augmented with past weather observations (solar irradiance and sun height) are compared against linear regression baselines. Results demonstrate that the multivariate TCN substantially outperforms linear regression at the one-hour horizon, and achieves consistent improvement at the fifteen-minute horizon. The relative contribution of architecture and weather features is resolution-dependent: at hourly granularity, the TCN architecture itself provides the dominant gain, while at the fifteen-minute scale the inclusion of weather covariates becomes the primary driver of accuracy, reflecting the greater atmospheric variability at finer temporal scales. A key finding is that past-only weather inputs are sufficient for accurate nowcasting, eliminating the need for future meteorological forecasts as model inputs. The results support the practical applicability of TCN-based models for national-level solar energy integration and provide a data-driven feature-selection criterion for similar renewable energy forecasting tasks.

Very short-term forecasting of solar photovoltaic energy production at national scale is challenging due to the high variability and spatial aggregation of generation across large territories. This paper evaluates Temporal Convolutional Networks (TCN) — a deep learning architecture based on causal and dilated convolutions — for nowcasting national-level solar production at one-hour and fifteen-minute horizons, using data from Spain sourced from the European Network of Transmission System Operators for Electricity. Multivariate models augmented with past weather observations (solar irradiance and sun height) are compared against linear regression baselines. Results demonstrate that the multivariate TCN substantially outperforms linear regression at the one-hour horizon, and achieves consistent improvement at the fifteen-minute horizon. The relative contribution of architecture and weather features is resolution-dependent: at hourly granularity, the TCN architecture itself provides the dominant gain, while at the fifteen-minute scale the inclusion of weather covariates becomes the primary driver of accuracy, reflecting the greater atmospheric variability at finer temporal scales. A key finding is that past-only weather inputs are sufficient for accurate nowcasting, eliminating the need for future meteorological forecasts as model inputs. The results support the practical applicability of TCN-based models for national-level solar energy integration and provide a data-driven feature-selection criterion for similar renewable energy forecasting tasks.