Sentiment analysis is something very new and immature. Sentiment analysis allows you to identify general attitudes towards your brand (or a product).
The goal of sentiment analysis is to assign a given social media comment with a degree of association to three basic categories: positive, negative or neutral.
1. Social Mention
Social mention is always my first choice when it comes to the sentiment analysis. It allows to search in multiple categories including:
- Blogs;
- Q&A sites;
- Twitter;
- Social bookmarking sites;
- Images;
- News, etc.
Socialmention’s sentiment is “the ratio of mentions that are generally positive to those that are generally negative.” They use textual analysis, emoticons, symbols to analyze the sentiment.
While many people call it flawed, SocialMention’s sentiment analysis is as good as you are currently able to get from a machine.
Socialmention’s sentiment ratio:
Socialmention’s sentiment distribution:
2. AlertRank Sentiment Analysis
AlertRank, a tool I reviewed just a few days ago, is a web based utility that adds some really great features to Google Alerts.
It allows you to configure the sentiment manually by clicking an icon next to each Google Alerts result:
After that the average sentiment is visualized via a nicely-looking graph:
Twitter Sentiment Analysis
With Twitter real-time search, sentiment analysis has become a particularly hot topic. Here are 4 free, registration-free services that can help you analyze Twitter sentiment of your brand:
How sentiment is analyzed | Limitations | Tracking | Visualization | |
Twitter Search | and icons used in a Tweet | none | RSS feed of search results | None |
Twitter Sentiment | Twitter sentiment classification (~80% accuracy) | Only one page of results is analyzed | Saved searches (you’ll be able to access your search online) | Graph |
Twitrratr | Analyzes the search term against a list of positive and negative keywords | No history | None | Color-based visualization (words that defined the sentiment are also highlighted) |
Twendz | combination of keywords and symbols + cross-reference against a list of positive and negative keywords | 70 recent Tweets | None | Yes |
(Also see screenshots of each tool below)