I am excited to be part of the “Innovators and Innovation” Panel at the Sentiment Analysis Symposium coming up May 8th. The event is a fascinating blend of top industry leaders and well known academics chaired by Seth Grimes of AltaPlana. The purpose of the panel is to lead discussions into where the field is going in terms of new innovations and uses. Many people are not even aware of what the uses of Sentiment Analysis are today, much less in the future. Even for many companies that are familiar with it, the technology is used to help them answer the question “is my brand healthy?” but they don’t explore ways that it can be useful outside of brand-building.
The term “sentiment analysis” itself is somewhat unfortunate as it brings to mind “feelings” and “opinions” implying that it only can be used where an opinion holder exists as the author of some sort of textual material. This is just one particular area covered by the technology. In fact, the opinion holder can be the reader of a particular text interpreting the set of facts described therein as “good” or “bad”. For example, sentiment analysis can be used on a corpus of news data to separate “good news” from “bad news”. This is a very worthwhile thing to do and very useful for business. The same technology could be also used to detect negative earnings reports in annually filed corporate documents and a host of other financial indicators. Furthermore, sentiment analysis does not need to be used only on documents but can be used on word groups as short as sentences or Tweets. So really “Sentiment Analysis” is an umbrella term that describes a set of classification techniques operating on a wide variety of texts.
I am eager to see where some of my fellow panelists and speakers go with their ideas. In particular, I am interested to hear the talk by Richard Brown of Thomson Reuters on Forecasting Financial Market Response and to hear Ryan Sager of the Wall Street Journal talk about how the media is using Sentiment Analysis. As for myself, I will be considering issues pertinent to our customers – how is sentiment analysis relevant to a job or candidate search? It is actually very relevant for both. As a jobseeker, you already know you should be researching the companies you want to work for. But most jobseekers interpret that as researching only the company’s products and services – instead of finding out what it would be like to work there. There are plenty of ways to find out how people feel about working at a company. Sites like Glassdoor have employee reviews and can help the jobseeker learn about the corporate culture. Another fun thing to do is check the Twitter stream for the company name. There is actually a nifty tool out now just for checking sentiment on Twitter called Sentiment140.com. I just tried it out using the search term “TheLadders”. Pretty cool!
So if jobseekers were to be sentiment-analysis-savvy, they would learn a great deal about the culture of the companies they are applying to. The same goes for recruiters. As a recruiter you need to be on top of your company’s reputation and know what employees are saying about it. It is much easier to close candidates if you can convince them that they will like the work environment and be able to back that up with evidence.
It looks like I will be the only participant from my industry at the symposium. I actually look forward to that since I will probably learn about new ways to use sentiment analysis and learn new data processing techniques that I can ultimately use right here at TheLadders!
Leslie Barrett is the Senior Search Architect at TheLadders. Leslie has worked making enterprise search software for companies large and small for many years. She holds a Ph.D. from New York University in Computational Linguistics, is a frequent speaker on issues in Search and Sentiment Analysis and is the author of over 20 academic papers on language technology.