« back to Projects
I’m Subbu Vincent. My interests are in helping people track their city without watching it. Research I did at the John S Knight fellowship at Stanford led to my conceiving of Civic Radar as an asynchronous alerting solution for journalists and civic activists, using open data and open documents. http://jsk.stanford.edu/journalism-challenges/2016/how-can-we-keep-city-newsrooms-in-the-loop/
For my data monitoring-and-alerting prototype, I am deliberately taking an off-the-shelf tool approach (with cloud options) to make it easier for non-technical people in newsrooms and non-profits to drop the barrier to adoption. Currently this is with Splunk. My use-case newsroom partner for this is the San Francisco Public Press.
For polling the targeted data sets every night, some customisation in Splunk is needed, and little python code, but that is common for the whole city’s datasets. From a technical configuration standpoint, once this is done the alerting solution can run on a cloud (hypothetical now, unfunded) with end-users across the city. Currently I am working with Splunk Enterprise installed on my laptop and on the Macs at SFPP’s office.
I have setup Splunk’s REST API input to poll the json API endpoints of several SFGov datasets. I currently need to write perhaps 10-20 lines of python code to implement a cursor - my polling loop each night has to pickup where it left off the previous poll. That way I will index only the updated records. I’ve been in touch with SFGov and Socrata folks to make sure I am calling the REST URL with the right system fields.
|Skills Needed||REST APIs python nosql|
|Tags||asynchronous alerts Splunk REST API data integration nosql python json csv big data socrata SODA|