I like word cloud, and have always wanted to make a word cloud that can describe myself. Recently I got laid off from Caterpillar. As normally done by many people, I updated my LinkedIn profile to reflect my most up-to-date work history. I also thought of asking recommendation from people (also like many people did). I noticed that most people would ask for their past bosses or managers, probably people think that those recommendations would carry more weight.
For me, I don’t really want to know just what the managers think of me. Frankly, I had managers that either not having LinkedIn profile, or not really keen in writing recommendations. So, I also started asking my peers. I think these recommendations are more valuable than from the managers, as I worked with them day-in and day-out. I’m lucky that I have collected quite a number of them, and I really appreciate those feedbacks!
OK, enough of the background of recommendations. I think with these recommendations, I can probably do some “analysis”. I came across this site (for analyzing Jeopardy games) and this site (more extensive tutorials), which I followed them through. For the steps in the Jeopardy games one, there is an issue that was solves in this site (basically do “tolower” to address issues with capitalized words).
I copied all my 20 recommendations (as of Aug 25, 2016) down to a file, then imported to R. Here is the lists of my “stopwords”:
stopnames <- c(“chinpei”, “tang”, “caterpillar”, “cie”, “chinese”, “engineers”, “students”)
stopplaces <- c(“dallas”)
stopmisc <- c(“his”, “her”, “person”, “they”, “them”, “student”, “year”,
“work”, “also”, “can”, “one”, “even”, “will”, “time”, “worked”)
stopverb <- c(“make”, “look”, “include”, “made”)
stopcombine <- c(stopwords(kind = “en”), stopnames, stopplaces, stopmisc)
My selection of stop words are basically the names, places, and the neutral nouns and verbs. There are probably more that are not really relevant, but I think these filtering are good enough, and not really filtering out the “bad stuff”.
Here is the word cloud that I generated:
Well, of course I have “skills”, and I’m in “technical” stuff, doing “engineering”, those terms are emphasized. I’m pleased to see words such as “pleasure”, “excellent”, “leadership”, “good”, “great”. In any case, let me plot a histogram of the occurrence of the terms:
Overall, thanks for all the peers who had faith in me providing me an insight that I can see myself. However, I am aware that those are the possible common terms used in recommendations in LinkedIn. For instance, here is what you can find about the 10 years of recommendation letters written for graduate school applications. Still, I think if a person doesn’t deserve a recommendation, he/she wouldn’t receive them. There is still some merit to it.