Reading the data turned out to be difficult; part of the group worked on that, while the others developed sonification strategies with really simple test data. Data reading was not finished on time, so we presented only the design sketches.

Our initial interest was comparing single points to the average year timeseries, scaling a year to maybe 2-4 seconds, so 10 years would be 20-40 seconds.

First we generated a really simple synthetic average rainfall curve (a single sine shape), then the synthetic rainfall curves made up of the average plus a faster, softer sine shape as the deviation, which is different for the two points:

The sound design was also very simple:

The two points observed are represented on the two stereo channels.

The current rainfall value for the day at each point is represented by the density of drop-like sounds in each channel. In the example 1, it moves from low to high back to low:

Rain_A1_LoMidHiMidLo.mp3

The deviation is expressed in volume: average is silent, more deviation is louder.

It is also marked by either a lower resonant pitch (for below average) and a higher resonant pitch (for above average). In the example 2, it begins on the left channel: much below, much above, little above; then in the right channel, much below, much above, little above average:

Rain_A2_deviaL_LoHiLess_R_LoHiLess.mp3

In the final example, the simulated data and their deviation from the simulated average are played thru 5 years: one hears the deviations (both sine shaped) begin in phase, go out of phase, and sync again; and density (mapped from rainfall) changes independently from that:

Rain_A3_2pnts_5yrs_deviationsLR.mp3