Although pharmaceutical companies are responsible for the manufacture of thousands of medications many patients take and depend on each day, there has been recent controversy over some of the industry’s standards and practices. This includes the influence these companies have when it comes to marketing and advertising to consumers. This can be investigated using a sentiment analysis based on a sample of consumers using Twitter data using the search term, “Big Pharma”.

The mean polarity is a negative value, which means that overall, most tweets were negative. Based on the minimum and maximum shown (which are also the minimum and maximum possible polarity values), however, it is clear that the 500 sample tweets collected cover the entire polarity spectrum. Selected words from across this spectrum can be seen in the wordcloud below.

Above is the same analysis run on the search term “Pharmacy”. As you can see, though the spread of polarities still span the entire spectrum, the tweets themselves were on average much more positive. This suggests that pharmacies themselves are looked upon more favorably than their suppliers. This is the trend I was especially interested in investigating. To start, I completed a geospatial analysis on the most recent 500 tweets for big pharma in order to find any “hot spots” that could provide some clues.

While there is a large tweet density around the northeast United States and regions of western Europe, it is important to understand that these are locations of certain manufacturers’ headquarters. Because people within a radius of these are more affected by the influence that the companies have on their local economy, it is expected that the densities of tweets there would be higher. Another high density “hot spot” seems to be surrounding Washington DC. This could be due to the fact that “Big Pharma” has been a topic most candidates have weighed in on, and like any political issue, people have VERY differing opinions, which could be expressed online.

Above are simple barplots representing the spread of the polarities of the gathered sample set of tweets. Just at first glance, the density of pharmacy polarities is much more top-heavy when compared to big pharma. So knowing that politics may in fact play a role in this disparity, the next question is to what extent does this occur? The first step is to look at the stock prices for the top community/retail pharmacy in America (Walgreens -WBA) and the top three drug manufacturers (Gilead Sciences - GILD, Johnson and Johnson - JNJ, and Pfizer - PFE). The span of time selected was two years so that within this range are election and non-election years for more accurate comparison.

These candlestick plots alone, however, do not provide enough information to posit whether a divergence between the stock prices for pharmacy and pharmaceutical companies has occurred now that we are in the thick of an election year. To do that, regression analyses must be completed.

The above chart features Walgreens and the top three manufacturers from April to the predicted values in June of this year. This can be used to measure the amount of correlation between the stock prices of these companies. We look to the adjusted r-squared value in particular, which in this case is a value of -0.1228. Below, the same analysis is run, but during the same two-month window in 2014, a non-election year.

The adjusted r-squared value for this chart is 0.6379. After taking the absolute value for each, the resulting difference is greater than half a point. This means that the stock prices of Walgreens, Gilead Sciences, Johnson and Johnson and Pfizer are significantly more related during years without a presidential election. Although further analysis incorporating more companies may supply more accurate results, this trend should be taken into consideration when working with a health-involved portfolio.

~ Emma Arents ~