Skills Requirements of Additive Manufacturing - A Textual Analysis of Job Postings Using Natural Language Processing

Additive Manufacturing (AM) has great potential of disrupting product design and supply chains in many industries by means of its unique capabilities when compared to traditional manufacturing. Moreover, the use of AM offers potential to produce in a more sustainable way. Despite these advantages of AM, its implementation still progresses rather slowly. Given the growing significance of AM, this research aims to identify the essential skills that an employee must possess to excel in this field and effectively utilize AM in the industry. For this purpose, this study is focused on using natural language processing algorithms to process data and analyse text from AM job postings. After gathering the information from the source sites using a web scraper and API, the data was cleaned, normalized, lemmatized, and tokenized. Word2Vec, LSTM and BERT algorithms were implemented to find the most requested skills. Using word2vec the algorithm was able to obtain similar words considering the inputted ones. Using the LSTM algorithm, the 35 most required skills were extracted. Further analysis was done, and the data was broken down by industry, AM technique, programming skills and design software. The third algorithm used was named entity recognition with BERT, which identified complementary skills to the ones mapped by the other algorithms. The three algorithms offered different and complementary insights to required skillset by AM industry, covering mostly hard and soft skills along the whole AM process chain. The use of each algorithm, corresponding user group and further developments are then suggested.