Anti-epileptic Drug Review Analysis

Yaxiong Huang
Posted on Feb 12, 2017

Introduction

Sometime when people have fever or headache or bad cold, they start to use tylenols.  People might ask which symptom that the tylenols take more effect on when comparing with other medicines.  In general, how do customers choose medications based on the reviews?  How can we answer these questions? These questions can be done by analyzing the drug reviews.  For this project, I am interested in analyzing the drug reviews for Epilepsy.  Exploratory data analysis, Sentiment analysis, and relation between rating and polarity analysis will be presented in the project.

Background

Epilepsy is a series of neurological disorders characterized by epileptic seizures and it is more common in older people.  Epileptic seizures are episodes that can vary from nearly undetectable to long periods of epileptic seizures.  The cause of most cases of epilepsy is unknown. Epilepsy can have both genetic and acquired causes.  Seizures are often brought on by factors such as stress, alcohol abuse, or a lack of sleep.   Nearly 80% of cases occur in the developing world.  Seizures  are controllable with medication in about 70% of cases.  What are the effectivenesses of the anti-epileptic drugs? How do they treat people with various symptoms differently? Does the drug take more effect on a particular symptom?  Pharmaceutical scientists might be interested in these questions so they can use the answers to do further improvement on drugs.  Also, how do reviews help customers in choosing drugs?  To answer these types of questions, we need to do an analysis on drug reviews.

Web Scraping

www.Drugs.com has drug reviews information.  For this project we selected 4 anti-epileptic drugs (GabapentinClonazepamTopiramatePregabalin) for the review analysis.

  • Gabapentin: It is used in adults to treat nerve pain
  • Clonazepam: It used to treat certain seizure disorders (including absence seizures or Lennox-Gastaut syndrome) in adults and children
  • Topiramate: Help to reduce seizure activity and prevent migraine headaches from occurring
  • Pregabalin: Slow down impulses in the brain that cause seizures

Below is a view on the data that we were interested in from the site.DrugReviewSnapshotEach review of the customer includes condition, review comment, and rating. In the snapshot above, the information will be collected are condition (For Reflex Sympathetic Dystrophy Syndrome: ), Comment ("I was diagnosed with RSD on Tuesday. ... I like these pills so far!"), and Rating (10).

The method for scraping the page is BeautifulSoup.

https://gist.github.com/tommyhuang1/4e99de23e7d82fd6b8f8f56563becfa9

Data Cleaning

The main item to be cleaned is the condition column.  Condition: "For Reflex Sympathetic Dystrophy Syndrome:".  The word 'For' in the front and ':' will be deleted.   Also, if the condition is "Neurontin (gabapentin) for Postherpetic Neuralgia", "Neurontin (gabapentin) for " will be deleted.

https://gist.github.com/tommyhuang1/53eeb3e87f5532631a9d1131fc975ac9

Below is the head of the data frame.

Screen Shot 2017-02-24 at 12.37.00 PM

Exploratory Data Analysis

Review Count By Condition

Gabapentin

Review_Count_By_Condition(Gabapentin)

https://gist.github.com/tommyhuang1/dd30de0dc41eedfa7a0853185b3c4c22

The majority of the reviews are from Pain and Anxiety for Gabapentin and follow by Fibromyalgia and Peripheral Neuropathy.

Clonazepam

ReviewCountByCondition(Clonazepam)

https://gist.github.com/tommyhuang1/06e0929c03c92e2d6664cc1877c59efe

The majority of reviews are from Anxiety and Panic Disorder.

Topiramate

ReviewCountByCondition(Topiramate)

https://gist.github.com/tommyhuang1/de70932a838adccc2127fa485d03ea64

Migraine has more reviews than other conditions.

Pregabalin

ReviewCountByCondition(Pregabalin)

https://gist.github.com/tommyhuang1/1a54e1251acf3921dff2fd585f79ea18

Fibromyalgia and Generalized Anxiety Disorder have more reviews.

Throughout this analysis we see which condition is used the most for the different drugs.

Average Rating By Condition

Gabapentin

AvgRatingByCondition(Gabapentin)

https://gist.github.com/tommyhuang1/776abfea93dc12ec0357c6b8fafe0752

Cough has the highest rating among other condition.

Clonazepam

AvgRatingByCondition(Clonazepam)

https://gist.github.com/tommyhuang1/0b5f6e30d8d578d791be52741039063e

Temporomandibular Joint Disorder, Chronic Myofascial Pin and Obsessive Compulsive Disorder have higher ratings.

Topiramate

AvgRatingByCondition(Topiramate)

https://gist.github.com/tommyhuang1/dab0361e958ae9bd8bcff9ac36deec3d

Seizures has the highest rating follows by Diabetic Peripheral Neuropathy and Tourette's Syndrome.

Pregabalin

AvgRatingByCondition(Pregabalin)

https://gist.github.com/tommyhuang1/02252c3bb654442bf336c86d7f5ad3ac

Occipital Neuralgia has the highest rating, follows by Restless Legs Syndrome and Dercum's Disease.

Throughout this drug review analysis, we found Gabapentin is good for cough. Clonazepam is good for Temporomandibular Joint Disorder, Chronic Myofascial Pin and Obsessive Compulsive Disorder.  Topiramate is good for Seizures and Pregabalin is good for Occipital Neuralgia.  The pharmaceutical companies could look at the lower average rating conditions and do further analysis on the drug ingredients.

Average Rating Comparison

AvgRatingComparison

https://gist.github.com/tommyhuang1/6333ef26bc419feecebfd9183fbcbbdd

Clonazepam has the highest average rating among the drugs and Topiramate has the least average rating.

Sentiment Analysis

Sentiment-Analysis

Review Polarity

ReviewPolarity

https://gist.github.com/tommyhuang1/e25572356da699bb5c9444c49cfce469

Clonazepam has the best polarity (more positive and less negative comments) among the other three drugs.

Review Subjectivity

ReviewSubjectivity

https://gist.github.com/tommyhuang1/1548739c63b45b69ac386d640da7fa3f

The review subjectivity scores are ranging from 0 to 1.  0 is being most objective and 1 is being most subjective.  All drugs have the same structure, it is almost normally distributed. There is a small amount of reviews being more objective.

 

Relation Between Rating and Polarity

Do the review comments match up with the ratings?  If there is a higher rating, do you think the comments will be positive? We will look at the scatter plots to make our conclusion.

GabapentinScatter

https://gist.github.com/tommyhuang1/4f8f7560ec94ab9b3ab3554b82e9b3c6

ClonazepamScatter

https://gist.github.com/tommyhuang1/b38b78a55907ccb95f12ff1db0934291

TopiramateScatter

https://gist.github.com/tommyhuang1/bf5c196177c3c86b439f5ac31e7abe6e

PregabalinScatter

https://gist.github.com/tommyhuang1/109e18c9e230f1facf141f01ad7485be

It is surprise that we do not see there is a strong correlation between rating and review polarity but we do see higher rating corresponds to higher positive score.  In the scatter plot of Gabapentin, there is a review with rating of 10 but there is a very negative review associated with it. The reason is the review mentions the customer was very ill before with 'miserable' on it and the other words are neutral.  The majority of the reviews for each rating are between -.5 and .5 polarity.  It is still quite normal as they talk about their illness in a negative way while mentioning about how do the drugs cure them in a positive way.

Conclusion

  • The scatter plot for rating and polarity does not reflect there is a positive correlation between them.  On average, the polarity is always a bit way from neutral in with positive or negative.  We do see the polarity scores are higher for the ratings that are above 7.
  • In terms of subjectivity, there is a small group of people with more objective  in writing reviews.  This makes the overall reviews be less bias.
  • Based on the sentiment analysis, customers should choose Clonazepam out of the 4 drugs for this case study because it has the highest positive score and lowest negative score.  In general, customers could compare more drugs by using this technique.

About Author

Yaxiong Huang

Yaxiong Huang

Tommy Huang received his Master of Arts in Statistics at Hunter College and Bachelor of Science in Mathematics and Economics at College of Staten Island. He has 7 years of experience in catastrophic modeling research for the insurance...
View all posts by Yaxiong Huang >

Related Articles

Leave a Comment

Avatar
Drug Education December 17, 2017
Hi there, thank you for the great information. I did have a question though I think you might be able to help me with. I was wondering, Can I lower my doses of gabapentin from 3600 to 900 a day without problems? I really want to decrease my dosage. If you could provide a little insight I would greatly appreciate it!
Avatar
Drug Education December 12, 2017
Hi there, thank you for the great information. I did have a question though I think you might be able to help me with. I was wondering, What do you personally feel when you take gabapentin for anxiety? If you could provide a little insight I would greatly appreciate it!
Avatar
running tips March 5, 2017
Great post.

View Posts by Categories


Our Recent Popular Posts


View Posts by Tags

#python #trainwithnycdsa 2019 airbnb Alex Baransky alumni Alumni Interview Alumni Reviews Alumni Spotlight alumni story Alumnus API Application artist aws beautiful soup Best Bootcamp Best Data Science 2019 Best Data Science Bootcamp Best Data Science Bootcamp 2020 Best Ranked Big Data Book Launch Book-Signing bootcamp Bootcamp Alumni Bootcamp Prep Bundles California Cancer Research capstone Career Career Day citibike clustering Coding Course Demo Course Report D3.js data Data Analyst data science Data Science Academy Data Science Bootcamp Data science jobs Data Science Reviews Data Scientist Data Scientist Jobs data visualization Deep Learning Demo Day Discount dplyr employer networking feature engineering Finance Financial Data Science Flask gbm Get Hired ggplot2 googleVis Hadoop higgs boson Hiring hiring partner events Hiring Partners Industry Experts Instructor Blog Instructor Interview Job Job Placement Jobs Jon Krohn JP Morgan Chase Kaggle Kickstarter lasso regression Lead Data Scienctist Lead Data Scientist leaflet linear regression Logistic Regression machine learning Maps matplotlib Medical Research Meet the team meetup Networking neural network Neural networks New Courses nlp NYC NYC Data Science nyc data science academy NYC Open Data NYCDSA NYCDSA Alumni Online Online Bootcamp Open Data painter pandas Part-time Portfolio Development prediction Prework Programming PwC python python machine learning python scrapy python web scraping python webscraping Python Workshop R R language R Programming R Shiny r studio R Visualization R Workshop R-bloggers random forest Ranking recommendation recommendation system regression Remote remote data science bootcamp Scrapy scrapy visualization seaborn Selenium sentiment analysis Shiny Shiny Dashboard Spark Special Special Summer Sports statistics streaming Student Interview Student Showcase SVM Switchup Tableau team TensorFlow Testimonial tf-idf Top Data Science Bootcamp twitter visualization web scraping Weekend Course What to expect word cloud word2vec XGBoost yelp