From tree to cloud, here are the words that define medical cannabis and tension user sentiments.
When we started researching medical cannabis there was too much confusing information and not enough science backed data. There are thousands of reviews for tension and medical cannabis, but it was difficult to make sense of it.
>> Enter data science. >>
TrueMedicines analyzed thousands of reviews for tension and medical cannabis and synthesized it into some pictures that the we can all understand.
Word Cloud for Tension
Interactive Word Tree for Tension
Data Science Approach
We collected and processed market research data in the form of consumer reviews. The datasets are often quite large, up to 200,000 reviews. It is imperative to identify the context within the reviews text so that we can incorporate these as features in our machine learning models. We have developed a medical corpus, or dictionary, of words to match reviews to medical conditions. Using this corpus we developed a medical matching algorithm to extract reviews with medical conditions only and tag these reviews with a binary value so that we can run statistical analysis on them to find reviews that are strongly correlated with the specific condition we are researching.
Next, we perform sentiment analysis and emotional analysis on the user reviews to derive a sentiment score from each review and classify reviews into three groups - positive, negative and neutral. A combination of sentiment score, emotions and user ratings provides a measure of efficacy.
About the Word Cloud
Keywords are extracted and visualized as a word cloud to make the research accessible and easy to understand.
About the Word Tree
We also visualize the results of text analysis in the form of a word tree that displays the connections between words. Font size of the text is proportional to the count, or the number of reviews which contain that particular medical condition or term. A word tree is also known as a dendrogram.
TrueMedicines is using market research and data analytics to discover the efficacy of plant-based remedies. We use data science, including machine learning and natural language processing, to research the efficacy of specific nutraceuticals for specific medical conditions.
Disclaimer: TrueMedicines research is informational and not personal medical advice. Our market research is not medical advice and is not intended diagnose, prevent, treat or cure any diseases. Contact your qualified physician for medical advice.