Nowcasting from a Different Perspective: A Literature Review with a Stack of Papers
A Literature Review from an Alternative Viewpoint
It is interesting to explore literature on this topic through text analysis, NLP, and quantitative approaches. The focus on these topics is also expanding to include qualitative approaches. I collected research papers from two sources: 48 papers from the Social Science Research Network (SSRN), which are the most downloaded papers with the keyword ‘nowcasting,’ and 43 papers from Google Scholar using the keywords ‘GDP nowcasting and Google Trends.’
Some technical tricks I used include removing stop words and defining specific steps for creating word clouds. For example, I separated all words that precede four-digit numbers in brackets, which allows us to see references or the most cited authors clearly. This way, we can identify which authors are frequently cited in these papers; if an author’s name appears larger in the word cloud, it indicates that they are cited more often on this topic.
Word clouds show the most cited authors in two groups of papers on nowcasting. The first group consists of the most downloaded SSRN papers with the keyword ‘nowcasting,’ which are primarily academic. The second group is from Google Scholar and includes more institutional working papers.
The observed trends for ‘google’ and ‘survey’ suggest that Google Trends data is increasingly recognized as a valuable tool for capturing economic sentiment and can sometimes serve as a substitute for traditional surveys. This highlights the evolving role of digital data sources in economic analysis. The second plot’s focus on methodological terms such as ‘OLS,’ ‘Lasso,’ and ‘ARIMA’ reflects ongoing advancements in econometric techniques, with these methods gaining prominence in economic forecasting. Meanwhile, the third plot’s examination of ‘forecasting’ and ‘nowcasting’ illustrates the growing emphasis on real-time prediction models to monitor economic conditions.”
In conclusion, this literature review through text mining and NLP techniques offers a unique perspective on the field of nowcasting and economic forecasting. By analyzing trends in commonly used keywords, cited authors, and methodological terms, we can gain a clearer understanding of the evolution and focus of research in this area. The contrasting characteristics of papers from SSRN, which are mostly academic, versus those from Google Scholar, which include more institutional insights, reveal the diverse approaches and applications in nowcasting research.
The use of word clouds and keyword frequency analysis provides an innovative way to visualize and understand trends in academic citations and methodological preferences. For instance, the prominence of keywords like ‘OLS,’ ‘Lasso,’ and ‘ARIMA’ in econometric methods reflects a movement toward more sophisticated and varied modeling techniques in economic forecasting. Additionally, the increased relevance of Google Trends data alongside traditional surveys underscores the expanding role of digital tools in capturing economic sentiment and real-time changes.
This approach to reviewing literature highlights the potential of text mining to reveal nuanced patterns in research focus, citation networks, and methodological evolution.
Future work in this area could benefit from a deeper exploration of thematic trends over time, allowing researchers to anticipate emerging methodologies and key contributors in the field. By continuing to leverage text analysis, we can enrich our understanding of how the field of economic forecasting is advancing and diversifying.
