In modern days, with the proliferation of social media and online blogs, city related issues are actively discussed by residents in open public forums on theweb. It has thus become imperative for city agencies to properly analyze these
information available in public domain towards e ective city planning. For example, a city agency may need to know which parts of the city has pollutionissue because of garbage and how is it impacting the people. Similarly, a city
transportation agency may want to get insights on whether the public transport services are not commensurate with the demand in any parts of the city. The trac department may want to know the spatio-temporal distribution of regular
trac problems and their potential causes. Insights on crime a ected areas may aid the police department to prioritize the most a ected areas accordingly. Existing practises typically involve manual surveys and subsequent analysis by
the city departments - and thus lack the incorporation of feedbacks from the masses. Furthermore, an automated analysis of the mass feedback to generate meaningful and actionable insights is paramount interest.
Such analysis based on information available from the masses in publicly available social media feeds and web sources can complement the existing sensor based data processing for smart-city planning in two ways: (i) it brings in the voice of residents while generating meaningful insights on city related issues for the agencies; and (ii) it further avoids the logistics challenges of sensor deployment in many cities (especially in the emerging economies). Additionally, sensor data processing techniques can bump into feasibility and scalability challenges in the emerging economies given the unstructured way of operation (e.g.,no lane discipline leads to infeasibility in lane violation detection using cameras/videos). Thus, analyzing what residents are discussing about can provide a better insight for the agencies to identify and prioritize the problems.

Generating meaningful knowledge from the public sources is non-trivial. First, the sources are not standardized and can be heterogeneous in nature with open information format. Second, an online solution is needed where millions of posts (reports) need to be collected, processed, and aggregated to identify distinct set of issues. Third, the issues need to be referenced to location landmarks in the city as much as possible for generating meaningful insight. For example, providing a garbage near the Total Mall in Marathalli, Bangalore is much more meaningful to the agencies than the precise location co-ordinates (i.e. latitude and longitude). Finally, a recommendation on the priority of the issues identifed for the agencies to handle is imperative. The problem further gets exacerbated by the need for an e ective interface, through which the aforementioned analysis on city related issues (and their priorities) can be accessed.