Adventures in Genealogical Data Science

Even since I started this website some 20+ years ago, I have in addition to putting up the family trees I have also provided databases of related information that I have found useful in my research and could be used by others, specifically the database of Churches and Priests extracted from Frank Perry’s book The Church in Madras (https://spuddybike.org.uk/research/madras-churches-and-priests/), and a database of those Born in India in the UK census’s in 1841, 1851 and 1861 (https://spuddybike.org.uk/research/born-in-india-in-the-uk-census-1841-1871/), extracted from public sources.

Over the last few years, there has been a bit of an explosion in relatively easy to use libraries for programming languages such as Python, that make nice charts, maps and other diagrams that help to visualise these trends. I have set up a GitHub (https://github.com/spuddybike/oharas-in-india) which contains anonymised extracts from the O’Hara family tree, and the code used to create charts and maps to try help me understand and explain the trends, and the bigger picture that emerges from over 1200 people in a 200+ year period covering 28 countries.

Many genealogy programs have the ability categorise people, put them on maps, create charts sometimes quite customised and sophisticated, but they don’t always help tell the story you want.

Births

Maps give you a sense of location in a way that it quite intuitive, this is a map of where O’Hara’s were born 1900-1920, the size of the circles give an idea of numbers and the locations, by putting, it into an animation it gives a better view of the changes over time.

A bar chart however, gives a better sense of scale, time and location at the same time and by just selecting a single descendant lineage, we can see that Amelia’s descendants moved to North America and then dispersed to other regions such as the UK and Australia.

Alternatively, the same plot, allows a comparison when put alongside all descendants

An alternate take is to use generations, this is a useful way of looking at the different trajectories in the case of the O’Haras where the births are very spread out, John’s children were born between 1830 and 1847, whilst George’s were between 1840 and 1895.

Births and Deaths

Perhaps, finding a way to add the death location, might give an insight. A sunburst chart seems like a good solution. Here we capture the birth and death region and isolate each by generation. Some close reading and patience (the programming libraries really struggle with good labels in the right location!) it’s possible to glean some useful information, but it is far from ideal.

If we simplify and remove the generational information, using a chord chart, this gives a different insight (albeit without time information).

Births, Marriages and Deaths

Sankey or Alluvial diagram are often used to look at flows in finance and other areas, perhaps this might be a useful way of combining them all together? This diagram below is for all people in the O’Hara tree who are not alive, basically anyone born before 1930. A good overview, but knowing the migration and being combined across all people, perhaps slightly limited.

Adding in generational information in this case just the four surviving children of Lawrence, not that interesting!

The second generation, starts to show something more interesting, marriage seems to be a proxy for intent to migrate, half of those born in India then get married and then move to another country.

Comparing Generation 2 with Generation 3 born outside of India they mostly stay in their parents country of birth, topped up by further migration from India and Burma

More recent generations, which are for obvious privacy reasons not publicly available can also be part of the story, genealogy is not just about the deceased!

Where next?

Are there other combinations, other charts or completely different ways of thinking about it …. I’d be interested in your thoughts and ideas, comments are open on this blog