LOESS (locally estimated scatterplot smoothing) is a method for producing a smooth curve through a set of scatterplot points. The implementation here is by Jason Davies from his library Science.js.
Adjust the slider to see how the curve changes depending upon its bandwidth. Higher bandwidths produce smoother lines, and vice versa. According to Wikipedia:
A user-specified input to the procedure called the “bandwidth” or “smoothing parameter” determines how much of the data is used to fit each local polynomial. The smoothing parameter, α, is the fraction of the total number n of data points that are used in each local fit.
The chart displays the average annual temperature in California from 1900 to 2017. Source: NOAA.
A prettier version of this chart was published in Axios on Nov. 15, 2018, under the headline “Humans are a wildfire threat multiplier”.