A heatmap with many columns has dozens of squares. An analyst wants a short list of the pairs worth a closer look: the ones that move together most strongly, whether up or down.
Task: write strongest_pairs(columns, k). columns is a dictionary mapping each column name to its list of numbers (all lists have the same length).
Return a list of up to k entries [name_a, name_b, r]:
name_a before name_b alphabetically. Never pair a column with itself.r is their correlation rounded to 2 decimal places.r, ignoring its sign, largest first. So -0.9 comes before 0.8.name_a, then by name_b.k pairs, return them all.There are at least two columns, no missing values, and no column holds the same value in every row.