Using a forked copy of Johns Hopkins data to load trends on COVID-19 infections into Neo4j.
This repo contains cypher scripts and shell scripts to import the data into Neo4j, which is a very small dataset at this point.
The original forked JHU data in raw CSV form can be found here
Simplified representation of what kind of data you end up with:
(:Region)<-[:IN]-(:Province)-[:REPORTED]->(:Report { confirmed, deaths, recovered, latitude, longitude, lastUpdate, reportDate }),
(:Report)-[:NEARBY]->(:City)-[:IN]->(:Country),
(:Region)-[:IS]->(:Country)
(:Province)-[:IS]->(:City)
- This repo assumes Neo4j 4.0! There were some subtle changes to cypher, things like toInt -> toInteger, and the load scripts probably won't run on 3.5
- You need APOC installed.
Nothing else
Check the import-all.sh
script. Adjust particulars for your situation, and then run it to re-create a database
from scratch.
./import-all.sh
This is a fast moving situation.
I've pointed this at a forked copy of the JHU data so that it can't go missing, but this necessitates keeping the fork in sync.
See "files-to-load.list" for a list of dated report files that will be loaded. This must be maintained to some extent as new data comes in.