Context Japan is a country with a unique history, and beautiful culture. In recent years, there have been concerns voiced about the size of Japan's population. The data set presented here will allow us to examine this issue, and see which regions of Japan are still growing in population, and which are not. It will also allow us to examine historical population trends in Japan, and map out how the population has been distributed at different points in history. Content Inside this data set you will find the raw population numbers for the prefectures, going as far back as the 1870s. Obtaining the oldest population data was challenging, as the late 1800s were a time of rapid growth and restructuring for Japan. There are currently 47 prefectures in Japan. However, in the 1870s there were almost double that value. I did my best to track down which prefectures merged into the modern prefectures (and when they did so), but my matches may not be perfect. This leads to a few of the population curves for the prefectures having "jumps". In addition, data for Okinawa prefecture from 1945 to 1948 could not be obtained. In addition to the population figures, the other columns have the name of the prefecture, its capital, the region of Japan it is a part of, and the island the prefecture is on. The estimated area of the prefecture (in km^2) is also available. Acknowledgements The data was obtained from Wikipedia, and the data on that site was obtained directly from the official census of Japan (and what preceded it, as the name and methodology of this census has changed over time). Inspiration This data set could benefit from the interpolation of missing population values. For example, as mentioned the population record for Okinawa prefecture is incomplete. Obtaining estimates of the population of this prefecture during this time period is a little complicated, since during this time the island was occupied by the United States military. Other prefectures have incomplete records for other reasons. A simple way to interpolate the missing values for these prefectures would be to assume a logistic growth model for each prefecture, fit such a model, and interpolate. However, there may be better ways of interpolating than this. It is my hope that this data set could also be paired with other data sets on Kaggle, and be of use to others.