Click here for a little background behind this little app

I had often opened the Wikipedia page for various cities and inspected the table of monthly temperature lows and highs to get a sense of what the climate was like. This app is a generalization of that.

The details are on GitHub but in a nutshell, I grabbed all ~13'500 global weather stations from the Global Historical Climatology Network daily (GHCNd) dataset that had temperature data over the last three years and looked at the temperature percentiles. I think percentiles are a more compact set of numbers to understand compared to the monthly averages Wikipedia has: the table below will have the specifics but a 50-percentile low means that half the days in the last three years were less than that value. A 10-percentile low means ~5 weeks of a year have lows lower than this, and a 90-percentile low means ~5 weeks per year have lows higher than this.

So something I enjoy doing is,

  1. first select a weather station near my home (you can either search OpenStreetMap or just zoom around the map) and clicking .
  2. Then find some other faraway place and click .

That should get you started! And sorry, as you can tell from the buttons and the widgets and the cobbled-together nature of everything, this was thrown together in literally a weekend, but I hope you like it!

(You might enjoy reading more about this app in my blog post about it. The name comes from Makoto Shinkai's film, Weathering with You, originally 『天気の子』, "weather child", where the eponymous child is called 「晴れ女」, "hareonna": a woman who brings good weather.)

▲ Close background

(Optional: Search for a place)

    Or: Click on a weather station and pick it

    Visualization of high/low temperature percentiles


    (Waiting for you to pick some weather stations.)

    Source on GitHub