Introduction
Text mining is a technique from data science approach that seeks distilling actionable insights from texts. This brief work presents an exercise of text mining and analytics based on the data available in the Monday Morning Manager (MMM) Blog Series of the Network for Social Work Management. The MMM is a weekly blog series, where social work managers referring to management skills and professional qualities, through sharing their experience in management successes and future challenges. Therefore; the goal was to retrieve relevant terms from the set of blog posts available in the NSWM website. This analysis was conducted using the RStudio software. To learn more about data science and social work read this paper which is part of the International Mentorship Program of the Network for Social Work Management final project.
The MMM is a weekly blog series, where social work managers referring to management skills and professional qualities, through sharing their experience in management successes and future challenges. Therefore; the goal was to retrieve relevant terms from the set of blog posts available in the NSWM website. This analysis was conducted using the RStudio software.
Methodology
The text mining process started retrieving and cleaning posts from the MMM blog series from June 2014 to February 2017. In each post, managers describe their experience on social work management path. In order to simplify the data retrieved and analysis, there were randomly selected four main features contained in 83 blog posts: leadership qualities, team motivation, networking implications, and advice to new managers.
This analysis used the “bag of words” method for text mining, where word type or order does not matter for analysis. Bag of words method focuses on the attributes of the documents. In other words, frequent words are the important key to proposing actionable insight.
In terms of technical procedures, the MMM blog post where organized as a “corpus”, which is a collection of documents, in order to reduce information and organize the text. Then, the corpus of the MMM blog post was cleaning by removing punctuation, stop words, and converting all the words to lowercase. In addition, the corpus was processed by a word steaming method to avoid similarities among terms and overrepresentation of specific words.
Results
Preliminary results of the text mining on the MMM blog posts were visualized by the word cloud method. This technique shows specific terms according to their frequency and compares them with other words contained in a text corpus. All the features were visualized using the word cloud technique.
Fig 1. Word Cloud for leadership qualities
In terms of social work qualities for leadership, the words more frequently used among mangers were “team”, “leader”, “staff”, “people”.
Fig 2. Word Cloud for team motivation
In this case, the most common word is “team”, which matches with the leadership qualities feature. In addition, terms such as “staff”, and “people”, are common as well. Words like “believe” and “know” are also frequent.
Fig 3.Word cloud for networking implications
In relation to networking implications, the word “people” is the most frequent term and match with the two previous features. In addition, words like “professional”, “relationships”, and “opportunities” are common within the text corpus.
Fig 4. Word Cloud for advice to new managers
For the final feature, again the word “people” is frequent. In addition, terms like “learn”, “know”, and “change” are common to refer advices to professional that are starting their career in social work management.
Fig 5. Pyramid plot for leadership qualities and team motivation feature comparison
This visualization shows the most common words expressed by manager related to leadership and motivation, as core components of social work management skills. Concepts and ideas like “team”, “staff”, “members”, and “people”, are predominant among managers.