Stack overflow surveys, and what they say about language trends
Since the 2022 Stack Overflow Survey is coming up, I thought it would be nice to look at the trends of past languages.
There are two metrics that are most useful on these surveys for gauging interest / popularity in programming languages, they are:
- Popularity of the language
- Love of the language
For the purposes of doing this article, I took the top 13 languages from the most recent survey (2021) that were not shell languages, database languages, or older/alternative versions of a language (like C).
Diagram 1 — Popularity
This is raw data for popularity provided by Stack Overflow.
Diagram 2 — Opinion
This is raw data from the love vs hate data provided by Stack Overflow.
Diagram 3 — Average Popularity weighted by Average Opinion
This is effectively a percentage of people who both used a language and loved that language over the three years: 2019–2021. This is why it is always less than the popularity. Ex: if 50% of respondents used a language and 50% loved it, that would show as 25%. From here on out we will be dealing with the data from 2019–2021 averaged together (or using it for growth %’s).
Diagram 4 — Popularity & Opinion Change Over Time
Diagram 5 — Taking Popularity & Opinion weight and applying growth and loss to simulate a future prediction.
If we take the data from diagram 3 and apply the growth and loss seen in diagram 4, we get a somewhat subjective “Outlook” value or a sense of where the language is headed in terms of overall popularity & love.
It should be noted that the long term outlook represents a prediction 9 years from now, and should be looked at as more momentum rather than a prediction. The shorter term “Outlook” is likely a better future predictor of where we will be in a few years, however even that can be quite far off. For example, I ran this same formula using older data (2017–2019) and got Java at close to 24% for 2022, this seems doubtful since it is currently under 20% with trends showing a decline.
Diagram 6 — Languages more loved than popular
These are languages that are more loved than they are popular, in some cases this could be considered “under-utilized,” “over-loved/hyped,” or it could be languages that are more specialized, newer, or still growing. To help account for this, the growth factor for both is included, so the following formula is used: (Love % * Love_Growth)/(Popularity % * Popularity_Growth )
Likewise, languages on the far right of this graph are more popular than they are loved.
Conclusion
I never loved the “what language is better” debate as every technology, principle, and concept has its place. If we were to attempt to generate a “best languages to learn” list it would make sense to weigh languages that were popular, of high opinion, and growing at the highest value. It is likely Diagram 5 would be the most useful in this measure. Below is the diagram with more data and presented in a different format.
The next developer survey is coming up for 2022! Once this data is in, I will do an update to this post.
