One major lesson I learned from this 6-year Ph.D. journey is, even without work-life balance, it is extremely important to have one’s daily life organized and automated. After spending some time using the reference manager Mendeley, I realized that some functionality like having a “watched folder” may be quite useful in the beginning of one’s research career, it could later on heavily reduce the signal-to-noise ratio in knowledge management since the literature collected tend to be falling into similar fields (even though the topic I work on is pretty interdisciplinary), which could produce many similar results when doing the search within the application using limited search terms.
A common knowledge if one uses Wikipedia often is, one terminology, if expressed in the same word/phrase in English, may represent different meanings based on the contexts. This could be time-consuming and sometimes misleading when the information is not well categorized. Instead of relying on the existing tools and pouring all one has into it, one should try testing a limited amount of data to examine the features s/he may need. For example, recently I tried to visualize my LinkedIn connection (~1500, April 2019), but due to the closure of LinkedIn API in the past few years, a very useful tool Socilab was down – I managed to find other APIs that are temporarily compliant to LinkedIn’s current policy – and help scrap data from my account. Before analyzing the downloaded csv file, and transporting all the profile screenshots to EndNote (powerful OCR and text search functionality), I hesitated. I don’t want the same thing to happen again (I right now have difficulty using my Mendeley since there are over 2500 files).
I am still contemplating on the next step.