Making a Graph DB ep4

In my previous article about my Graph Database, I found a list of free & open-source libraries for Elixir, Python and Node, and I had a homework: to find 3 a minimum of use-cases that I want to resolve with my Graph Database.

I thought about this a lot.

First use-case:

• add all the countries and their neighbors from Mledoze/Countries, the geo coordinates, all the capitals, all the currencies and the most important cities
• that’s a small graph, probably around 1k-2k nodes and edges, but is useful, because I want to see how I would replace a classic document-based DB with a graph
• be able to query by continent, by country, by capital, by currency
• find top 3 largest and smallest countries
• find top 3 countries with the most neighbors
• what are the top 3 most popular currencies (by country)
• stupid thinks like: what are the cities with the same name, from different countries
• are there any countries with the same capital? - sanity check
• what are the countries that span on more continents? - don’t even know if I have this data

Second use-case:

• add all my favorite movies
• that’s also a small graph, around 1k-2k nodes and edges, but is useful because the relations are many-to-many
• what period of time is the most dense with favorite movies
• what are my favorite actors
• what are my favorite genres

Third use-case:

• add the top 100 most popular Node.js libraries, with their dependencies
• I have no idea how large this graph is, but it’s definitely something pretty large and dense; if not I could add the top 1000, or top 5000
• find top 3 libraries with most and least dependencies
• find top 3 libraries with most and least versions

Forth use-case:

• this is just a stress test: add all english words from Wordnet-DB by type
• this is a medium graph, around 1 million nodes
• what’s the category with the most words

Other notable use-cases:

• lenght converters (eg: meter to km, meter to foot, meter to mile, etc.)
• weight, temperature, time converters as graphs

With that in mind, I can start studying the popular graph libraries that I found. This might take a lot of time…

Hope to see you again!