Some programming languages are better suited for certain tasks than others. In order to further explore this phenomenon, Quinn wishes to implement a trending language feature for POTW. This feature should determine a single trending language for each week. In order to capture the true essence of the word trending, the language selected should exhibit a spike in usage for that given week.
For example, assume that Python accounts for 60% of all POTW submissions (this percentage will be herein be referred to as the global mean). If for some week Python appears in 62% of submissions (the local mean), then it shouldn't be regarded as trending as this is only a slightly above average turnout. Conversely if Go's global mean is 10%, but it displays a local mean of 30% for a given week, this implies a spike in the expected number of Go submissions by a factor of 3. Hence, Go should be considered trending.
More formally, a language's trending score for a certain week can be expressed as the ratio between it's local mean and global mean. The language which renders the highest trending score for the week is thus crowned the trending language.
A new API endpoint has been added to the POTW website which exposes statistics on the programming languages used to solve the problems each week. This endpoint is located at: http://potw.quinnftw.com/api/solution_languages. Each submission is represented here as a (week, language) pair.
Utilize this new endpoint to dynamically determine the trending language for each week. The output format is up to you (brownie points will be awarded if you can integrate your solution into the actual website).