2020 was a memorable year where the world, metaphorically, stopped revolving due to the COVID-19 pandemic.
Now that the year has passed, market participants are eager to learn what the focus themes were for public companies and their aggregate impact from a sector perspective. From these themes, one can potentially form an eagle-eye view of the year 2020 within the sector of interest.
To identify these themes we opted to go with the straightforward method of Wordcloud, “an electronic image that shows words used in a particular piece of electronic text or series of texts”.  The words are different sizes according to how often they are used in the text.
The Wordclouds were created with data from S&P Global Market Intelligence’s Earnings Call Transcripts using the following approach:
- Limited the universe to S&P 500 constituents for the whole year of 2020, grouped by GICS Sector.
- Applied stop words such as “a”, “we”, “of” etc. in order to naturally highlight the words that best represent the themes and characteristics of the 11 GICS sectors.
- Constructed the Wordcloud utilizing the open-source Python Wordcloud package
Based on empirical evidence presented by the Wordcloud images, we can identify some of the key words within a sector. For example, words such as “rent” and “tenant” were specific to Real Estate, while “cost” seemed to be a common focus across Energy, Materials, and Industrials.
Surprisingly “covid” or “coronavirus” was not a word of focus across the sectors except for Healthcare. Instead “growth” was a more common focus word in a lot of the sectors.
This could mean that the companies were downplaying the impact of the pandemic to their businesses, or that they were steering the market’s attention towards growth and the future.
Unfortunately, Wordcloud cannot show you these advanced features explicitly. If one is interested to get a deeper interpretation between the lines, one needs to read the transcripts themselves or apply advanced Natural Language Processing (NLP) to the text for feature extractions and sentiment analysis. A NLP primer paper published by the S&P Global Market Intelligence’s Quantamental Research team describes this in more detail, and is a good starting point to further your NLP journey.
For those who are trained in the subject, looking at the Wordcloud itself may be enough to draw a picture of what messages and themes the companies were trying to convey through the earnings calls.
For those who are just getting started, can you also see and decipher any hidden themes in the Wordcloud?
Data enthusiasts can explore premium fundamental and alternative datasets available seamlessly via Cloud, Feed and API solutions, along with expert analysis on the new S&P Global Marketplace platform.
All the Wordclouds in this article were sourced from GitHub. The author wrote the Python code and the images are generated based on the code utilizing the Wordcloud package.
 WORD CLOUD: Meaning in the Cambridge English Dictionary. (n.d.). Retrieved January 22, 2021, from https://dictionary.cambridge.org/dictionary/english/word-cloud
 Machine-readable format of the Earnings Call Transcripts (available from S&P Global Market Intelligence’s flagship Xpressfeed) was utilized for this exercise
 Andreas Mueller’s Github Page Amueller. (n.d.). Amueller/word_cloud. Retrieved January 25, 2021, from https://github.com/amueller/word_cloud
Learn more about Market IntelligenceRequest Demo
Learn more about Marketplace
Natural Language Processing, Part I: Primer
Natural Language Processing – Part II: Stock Selection