Transcript of "Using Sentiment Analysis in Pharma and Biotech"
Sally Church, Ph.D
Understanding about how people feel about
their treatment and disease is critical
• PR tools give a big picture from media hits, but are less
useful for more granular analysis.
• Current buzz metric tools are useful for relative market
share of noise, but not for deep analysis.
• Traditional qualitative market research has limitations in
• Quantitative analysis for market segmentation is
expensive and usually based on physician rather than
• Rise of science, patient bloggers and social media as an
important mediums for evaluating sentiments means we
need new and better tools.
What is the basis of our approach?
• The underlying philosophy of the Leximancer
system is that meaning must be emergent from
text and therefore not biased by external pre-
• This has resulted in a methodology involving
identification of a Thesaurus of supporting
defining terms being 'bootstrapped' to an original
emergent seed-word, to form a Concept entity.
• It is a fully-automated discovery process using
the original text corpus.
Analysis can be from private recorded
interviews or public information on internet
Control panels allow a deeper analytical
framework to be applied by our analysts
Using a Thesaurus of words, +ve or -ve
associations can be established
The tool groups together word associations
to build a Concept Map
You can also drill down to specifics to look at
associations more carefully
We are now working with Pharma clients
using the tool in a number of different ways
• Qualitative interviews
Sentiments can be:
• Unmet medical needs • Therapies
PR and Advocacy: • Companies
• Media Monitoring • Physicians or Patients
• Issues Management
• Social Media Outreach