abehandler / rookie Goto Github PK

http://thelensnola.org/2013/10/14/live-blog-tuesday-mayor-mitch-landrieu-presents-2014-city-budget/

query mitch landrieu. "city council" not highlighted. "City council" is.

top graphs are squished.

q= [[british petroleum]]

'Mayor Mitch Landrieu', 'Mayor Mitch Landrieus'

right now each are an item in the aliases list

i think this is coming from the merge

if one name is last and the other is first,last then resolve to first last. same for acronyms

@myersnews stripping out all the punctuation makes it hard to read. If we’re at the point where we are trying to see if this makes sense to users, it’s time to add that in. Same with capitalizations. It’s like reading a stream of consciousness.

give the calendar value, not the string of the date

http://54.191.10.239/testing?q=Mitch%20landrieu

Marlin Gusman -- Sheriff, New Orleans, oversees Orleans Parish Prison

"Choice Foundation". they get selected together so not sure what's going on

not sure what's up

http://thelensnola.org/2011/01/27/independent-schoolhouse-authority

check

on #4: if both entities show up in the same sentence within an article, i think it should show the highest sentence under the headline

Ex. Search term: FEMA
Entity: Nagin

Katrina fact-check: Guesstimate of Katrina’s flooding in New Orleans was correct
In the interview, Nagin said he got the information from a “briefing that I got from a guy named Marty from FEMA.”

filter out low pmi

https://s3-us-west-2.amazonaws.com/rookielens/results.html?term=academic%20goals

Build a classifier for people?

http://54.191.10.239/testing?q=Mitch%20landrieu

"The policies"

is confusing

errors

I think you need to be able to drill down to the original text

print [j for j in twograms if j[0]=='Disciplinary Counsel']

search mitch landrieu --> narrow to consent decree --> narrow to splc and gusman.

that could happen w/in the ui results

http://localhost:5000/rookie/search?q=OPSB

I think a character encoding issue

https://s3-us-west-2.amazonaws.com/rookielens/results.html?term=Marlin%20Gusman

For instance:

'MONITOR SUSAN HUTSON' / 'POLICE MONITOR'

are not in both the count and the joint count. What is that about? Are the counts working correctly? A bug?

merger = KeyMerge([["President", "Barack", "Obama"], ["Senator", "Barack", "Obama"]])
print merger.compute_jaccard_index(set(["President", "Barack", "Obama"]), set(["Senator", "Barack", "Obama"]))

Delete bylines before processing

do a second round after elastic search to find the most important headlines

why are these showing up? they should not be

jail complex
to be home to the Templeman I and Templeman II jail buildings.It is not clear whether Gusman intentionally misled the working group

http://54.191.10.239/testing?q=coastal%20restoration

Southeast Louisiana Flood Protection Authority