What’s your opinion about…?
Needless to be asked, consumers have already expressed themselves in blogs and social networks. reviews, ratings, recommendations…they have shared their feelings about yours or other brands: someone loves iPod, others feel “abc” mp3 player sucks. Nowadays, online opinions can make or break a product in the market place. Sentiment analysis field is emerging.
Sentiment analysis or opinion mining refers to a broad (definitionally challenged) area of natural language processing, computational linguistics and text mining. Generally speaking, it aims to determine the attitude of a speaker or a writer with respect to some topic. The attitude may be their judgment or evaluation (see appraisal theory), their affective state (that is to say, the emotional state of the author when writing) or the intended emotional communication (that is to say, the emotional effect the author wishes to have on the reader).
I’ve previously introduced tools you can use to identify the online trends and track online conversations. Today, I’m going to share with you several free online tools for simple sentiment analysis to pinpoint the effect of specific issues on your customer perceptions, helping you respond with appropriate marketing and PR strategies.
Sentiment Analysis Tools Based on Twitter Tweets
Tweetfeel scours Twitter for Tweets about brands of your choice and shows you how positively or negatively Twitter users feel about it. The screenshot above shows results (72% positive) for “iphone” and the results show you a sample of recent tweets as well as an overall rating for the brand. It’s a quick way to get a feel for what people are thinking about your brand/product.
In sentiment analysis, you need to refine the keywords to identify the customers’ perceptions about different issues. Take the above analysis on “iphone” for example; we see even more positive results on “iphone 3gs” but very negative results on “battery” related issues:
Twendz. It’s a Twitter mining web application that utilizes the power of Twitter Search, highlighting conversation themes and sentiment of the tweets that talk about topics you are interested in (such as your brands). This tool seems more promising than Tweetfeel as you can directly join the conversation by clicking on the “Reply” button.
You can select how fast you want Twendz to update the results; you can see clearly the sentiment in the color bars on the left; and, a tag cloud of related terms is displayed.
Twitrratr searches Twitter for a keyword (your brand or product) and the results it gets back are cross referenced against its adjective lists, then displayed accordingly.
Sentiment Analysis Based on Online News
Newssift (a product from Financial Times) indexes content from major news and business sources and annotates all of this content (and excludes content that lacks credibility or business relevance). It matches business topics to your query (e.g. brands, legal risks, the launch of a new product, environmental impact) so that you can know about changing issues across time for a company or product.
Newssift allows you to build a complex query; the more specific your query, the more refined your results. After you have chosen a term in the query box, the tool will automatically move the box to the left and open up a new box to allow you add a term to refine your search.
Newssift suggestions are linked to the meaning and the relationship of a term or terms. These suggestions reflect the top people, places and things in the thousands of business news sources and million of articles they annotate for meaning, relationship and sentiment. You can click on to get further refined results.
If you are consistently researching a particular brand or business topic, you can have your searches saved. In future releases, you will also be able to set up alerts to let you know when breaking news occurs on topics you are researching. This feature will be very helpful to marketers.
Mood Analysis based on LiveJournal Posts
MoodViews is a collection of tools for tracking the stream of mood-annotated text made available by LiveJournal. At present, MoodViews consists of three components, each offering a different view of global mood levels, the aggregate across all postings of the various moods:
- Moodspotter lets you search for mood-associations in a period: you select the period and type in a word or phrase, and Moodspotter spots the top moods associated with this word or phrase in the selected period for you. (It doesn’t seem to be working when I tried this so I can’t tell how good this is.)
- Moodgrapher tracks the global mood levels, you can select the mood level (sad, sick, worried,…) with date interval; here is a sample for what many bloggers felt around Valentine’s Day 2006, while there were increased levels of flirty-ness and loneliness:
- Moodteller predicts mood. This seems a useful tool but it just didn’t work when I tried.
LiveJournal enables its bloggers to tag their posts with mood indicators. Moodteller analyzes about 5000 LiveJournal blog posts per hour and estimates, according to the textual features of the posts, the percentage of them which are “happy”, “sad”, “excited”, and so on. Moodteller does not use the mood indications given by LiveJournal. After the estimation is done, we check how good it is by examining the real indications given by bloggers.
- Moodsignals helps in understanding the underlying reasons for mood changes.
Semantic search engine Evri released a new sentiment web API which allows you deeper insight into the “who’s,” and “what’s,” and “why’s” associated with the particular expression or feeling. Find out more here. A few more paid sentiment analysis tools if you are interested:
Sentiment analysis allows you to identify attitudes to individuals and organizations and track key trends. Pinpoint the effect of specific issues on your customer perceptions, and respond with appropriate marketing and PR strategies.
Update (Sep 11, 2009): RankSpeed is a search tool that does a sentiment analysis on the blogosphere / twittersphere. You can search for any website category using tags and rank by any criteria: good, useful, easy, secure, etc…here’s a video on how to use RankSpeed.
You can get the results by sentiments: