Hey guys. I have a table with thousands of bus travel records, telling who was on the buses. What would you do at Siren to find out people who traveled more than once with suspect X?
only if I could group the documents by person (after filtering bus travels by the suspect, using the relationship person -> bus.)
I assume you have a dashboard created on table where each row has the structure
“person ID, bus busrideID” (busride might be a combination of busnumber+time of the bus e.g. a bus that is 411 might leave at 09:05 09:45 and then there are dates like 10022019 etc… so busrideID might be 411090510022019 as a single “token” (keyword)
PERSON ID, BusrideID
on this dashboard put:
- a term aggregation visualization where you see the “top busrideID”
- another one that gives “top persons”
now put a filter for your suspect.
this will give you just the rows where your suspect took a ride.
Now look at the “top bus rides” visualization… obviously you’ll only have “1” so that will pretty much look like the list of bus rides your suspect took. Good.
Now keep CTRL pressed and click on ALL those rides… you’ll get a big OR filter with all the bus rides together. e.g. 411090510022019 411090510023019 411090510024019 .
Now remove the original filter/search for your suspect… this will give you all the rows of all the passenger for just the rides that are in the filter.
… look at the TOP PASSENGERS visualizations. The first one will obviously be your person (the suspect) but then the other will be the top ones that have been with him/her… bingo
If you don’t have clear passeger IDs but instead you have fuzzy names, incomplete… e.g. Giovanni … Geovanni…, addresses that dont match (e.g. passegers online registrations) you’ll need to use Siren ER first. (entity resolution engine) to clean that up… then do the analysis above.
Makes sense? Feel free to contact us for a chat, happy to explain in a call and there is more that can be done.
Oh its a good idea! thank u